[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-tabular-optimization-ingestion":3,"mdc--finq57-key":43,"related-repo-nvidia-tabular-optimization-ingestion":1949,"related-org-nvidia-tabular-optimization-ingestion":2039},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":38,"sourceUrl":41,"mdContent":42},"tabular-optimization-ingestion","infer optimization structures from tabular data","Infer optimization structure from uploaded tables and identify minimal clarifications before cuOpt modeling.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"Data Analysis","data-analysis","tag",{"name":17,"slug":18,"type":15},"Optimization","optimization",{"name":20,"slug":21,"type":15},"Data Modeling","data-modeling",{"name":9,"slug":8,"type":15},462,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples","2026-07-30T05:29:20.240418","Apache-2.0",81,[29,30,31,32,33,18,34,35,36,37],"combinatorial-optimization","gpu","linear-programming","mixed-integer-programming","operations-research","optimization-algorithms","route-optimization","traveling-salesman-problem","vehicle-routing-problem",{"repoUrl":24,"stars":23,"forks":27,"topics":39,"description":40},[29,30,31,32,33,18,34,35,36,37],"NVIDIA cuOpt examples for decision optimization","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fcuopt-examples\u002Ftree\u002FHEAD\u002Fcuopt_on_nemoclaw\u002Fopenclaw-skills\u002Ftabular-optimization-ingestion","---\nname: tabular-optimization-ingestion\nversion: \"26.06.01\"\ndescription: Infer optimization structure from uploaded tables and identify minimal clarifications before cuOpt modeling.\nlicense: Apache-2.0\nmetadata:\n  author: NVIDIA cuOpt Team\n  tags:\n    - cuopt\n    - nemoclaw\n    - orchestration\norigin: skill-evolution\n---\n\n# Tabular Optimization Ingestion\n\nUse this skill when the user provides raw or semi-structured data and asks a question that may require optimization.\n\nThe purpose of this skill is to bridge the gap between messy uploaded data and solver-ready model construction.\n\nThis skill does **not** solve the optimization problem itself. It inspects the data, infers likely modeling roles, and identifies what still needs clarification.\n\n**It does not authorize heuristic, greedy, or backtracking schedules as\nanswers.** In the NemoClaw sandbox, read `cuopt-sandbox`: the first solver\nthat produces assignments or a schedule must be cuOpt after probe → env →\nsmoke gates pass. Ingestion output is a modeling interpretation (entities,\nobjective fields, constraints) — never a completed plan.\n\nThis skill refines the optimization interpretation using the uploaded data; it does not replace the earlier intent decision unless the data clearly contradicts it.\n\n## Purpose\n\nUsers do not upload:\n- objective vectors\n- sparse matrices\n- explicit decision variable definitions\n\nThey upload things like:\n- products.csv\n- capacity.xlsx\n- orders.json\n- travel_times.csv\n- dealers.csv\n- depots.csv\n\nThis skill turns raw tables into a candidate optimization interpretation.\n\n## What this skill should produce\n\nAfter inspection, produce a compact working interpretation containing:\n- likely problem type: LP \u002F MILP \u002F QP \u002F routing \u002F unclear\n- likely entities: products, resources, orders, locations, vehicles, customers, depots, time periods, etc.\n- likely objective fields: profit, cost, margin, time, risk, distance, penalty\n- likely constraint fields: capacity, demand, budget, inventory, hours, availability, limits, min\u002Fmax bounds\n- likely decision grain: per product, per route, per order, per plant, per vehicle-stop, etc.\n- minimum unresolved ambiguities that require a user clarification\n\n## Ingestion workflow\n\n### 1. Identify table shape and grain\nFor each uploaded source, determine:\n- what one row appears to represent\n- whether the table is entity-level, transaction-level, matrix-like, or parameter-like\n\nExamples:\n- one row per product\n- one row per plant\n- one row per customer order\n- one row per location pair\n- one row per vehicle\n- one row per time period\n\nAlways state the candidate row meaning before modeling from it.\n\n### 2. Identify likely optimization roles by column meaning\nInspect column names and values for roles such as:\n\nObjective-like fields:\n- profit\n- revenue\n- margin\n- cost\n- travel_time\n- distance\n- risk\n- penalty\n\nConstraint-like fields:\n- capacity\n- demand\n- available_hours\n- budget\n- max_load\n- min_order\n- inventory\n- time_window_start\n- time_window_end\n\nIdentifier \u002F relationship fields:\n- product_id\n- plant_id\n- order_id\n- customer_id\n- vehicle_id\n- depot_id\n- location_id\n- origin \u002F destination\n- pickup \u002F delivery\n\n### 3. Infer candidate decision structure\nInfer what the model is probably deciding.\n\nExamples:\n- units to produce by product\n- units to ship from origin to destination\n- whether to open a facility\n- which vehicle serves which stops\n- route sequence through locations\n- allocation of budget across campaigns\n\nDo not overcommit when the data supports multiple plausible decisions. Note the candidates and ask one focused question if needed.\n\n### 4. Match the data pattern to a problem family\nUse common table patterns:\n\n#### LP \u002F MILP patterns\nTypical signs:\n- product\u002Fresource tables\n- profit and resource consumption by product\n- capacity tables\n- demand tables\n- assignment tables\n- open\u002Fclose or yes\u002Fno columns implying integrality\n\n#### QP patterns\nTypical signs:\n- covariance matrix\n- risk matrix\n- variance terms\n- squared error context\n- portfolio weights or regression-style structure\n\n#### Routing patterns\nTypical signs:\n- locations \u002F coordinates\n- travel time or distance matrix\n- depot + customer tables\n- vehicle tables\n- demands and capacities\n- time windows\n- pickup and delivery pairs\n\n### 5. Identify only the critical ambiguities\nBefore handing off, identify the smallest set of unanswered questions that block valid model construction.\n\nExamples:\n- Are demands mandatory or forecast-only?\n- Must decisions be integers?\n- Are these time windows hard constraints?\n- Can unmet demand be allowed with penalty?\n- Is profit net profit or revenue only?\n\nDo **not** ask broad generic questions if the data already strongly suggests the answer.\n\n## Preferred output format\n\nSummarize findings compactly:\n\n- **Candidate problem family:** ...\n- **Row meanings:** ...\n- **Likely decision:** ...\n- **Likely objective:** ...\n- **Likely constraints:** ...\n- **Unresolved blockers:** ...\n\nThis summary should be short enough that a downstream formulation skill can use it directly.\n\n## Column-role heuristics\n\nUse these heuristics carefully. They guide interpretation but do not prove it.\n\n### Cost \u002F objective heuristics\n- `cost`, `unit_cost`, `shipping_cost`, `expense` → likely minimization coefficient\n- `profit`, `margin`, `contribution` → likely maximization coefficient\n- `distance`, `travel_time`, `duration` → likely routing objective term or service constraint\n- `risk`, `variance`, `covariance` → likely QP signal\n\n### Capacity \u002F demand heuristics\n- `capacity`, `available`, `limit`, `max_*` → upper bound \u002F resource constraint\n- `demand`, `required`, `need` → demand fulfillment or service requirement\n- `min_*`, `minimum_*` → lower bound or service rule\n\n### Integrality heuristics\nTreat MILP as likely when the data or request suggests:\n- whole units\n- counts of vehicles, workers, facilities, shifts, items\n- yes\u002Fno choices\n- on\u002Foff decisions\n- assignment indicators\n\n### Routing heuristics\nTreat routing as likely when the core question depends on **path construction**, not merely allocation.\n\nSigns include:\n- stop sequence matters\n- travel between locations matters\n- vehicles start\u002Fend at depots\n- travel matrix or coordinates are present\n- pickup and delivery relationships appear\n\n### Cross-row coupling heuristics\n\nDuplicate values in a foreign-key column across rows of a parent table usually signal a **shared agent or resource** — one entity serving multiple parents. Shared resources need a mutual-exclusion constraint that no other column states.\n\n- `coach_id`, `driver_id`, `instructor_id`, `nurse_id`, `operator_id` → shared agent; can serve only one parent at a time\n- `machine_id`, `bay_id`, `tool_id` → shared resource; can host only one job at a time\n- any FK column where `distinct_values \u003C row_count` → check whether simultaneous assignment is allowed\n\nA `*_unavailability` (or `*_availability`) table documents *known absences*; duplicated FK values document *implicit conflicts*. Treat both as constraint sources. Concrete check: for every FK-looking column in a parent table, compare distinct value count to row count, and surface the column when `distinct \u003C rows`.\n\n### Pre-solve capacity arithmetic\n\nBefore handing off to formulation, compare aggregate requirements against\nupper bounds implied by the tables:\n\n- required events or appearances per entity (e.g. round-robin games per team)\n- maximum placements allowed by per-week, per-day, or per-slot caps\n- total court\u002Fslot\u002Fmachine capacity across the horizon\n\nIf required appearances exceed what the time structure can host, flag the\nconflict before building the full MILP. Example: 8 required games per team\nwith at most 1 game per team per week over 5 weeks allows only 5 games —\ninfeasible without relaxing the weekly cap.\n\n## Examples\n\n### Example 1: product mix tables\nFiles include:\n- `products.csv` with columns like `product`, `profit`, `labor_hours`, `steel_units`\n- `capacity.csv` with columns like `resource`, `available`\n\nLikely interpretation:\n- one row in `products.csv` = one product\n- one row in `capacity.csv` = one resource limit\n- decision = how much of each product to produce\n- objective = maximize profit\n- constraints = labor and steel capacities\n\n### Example 2: routing tables\nFiles include:\n- `customers.csv` with `customer_id`, `demand`, `time_window_start`, `time_window_end`\n- `vehicles.csv` with `vehicle_id`, `capacity`\n- `travel_times.csv` with origin\u002Fdestination or matrix-style travel times\n\nLikely interpretation:\n- customer rows are stops with demand and optional time windows\n- vehicle rows define fleet capacity\n- travel table defines movement cost\u002Ftime\n- likely problem family = routing\n\n### Example 3: time-slot \u002F resource assignment (scheduling MILP)\n\nFiles include patterns such as:\n- `games.csv` or `jobs.csv` — items to place (events, tasks, orders)\n- `time_slots.csv` or `shifts.csv` — when placement can occur\n- `courts.csv`, `machines.csv`, or `rooms.csv` — resources\n- `teams.csv` or `workers.csv` — entities tied to shared agents (coaches, operators)\n- `*_unavailability.csv` — blocked (resource, slot) or (agent, slot) pairs\n\nLikely interpretation:\n- decision = assign each item to a (slot, resource) or similar binary\u002Finteger placement\n- hard constraints = no double-booking, unavailability, capacity, one game per team per slot\n- likely problem family = **MILP** (even if user only says \"build a schedule\" or \"valid plan\")\n- **NemoClaw:** read `optimization-from-data-orchestrator` + `cuopt-sandbox` before any custom scheduler code\n\n### Example 4: historical transaction table\nFile includes:\n- `sales_history.csv` with `order_id`, `date`, `region`, `revenue`, `units_sold`\n\nLikely interpretation:\n- this may be analytics, not optimization, unless the user asks for a future decision under constraints\n- do not invent decision variables from history alone without a decision-focused question\n\n## Guardrails\n\n- Do not assume every table with `cost` and `capacity` is automatically a valid optimization model.\n- Do not mistake transaction history for decision variables without evidence.\n- Do not treat forecast data as hard constraints unless the user implies that.\n- Do not infer QP unless there is a real quadratic signal.\n- Do not infer routing unless movement between locations is central to the decision.\n- Prefer one precise clarification over a long list of speculative questions.\n\n## Handoff guidance\n\n- If the data suggests LP \u002F MILP:\n  - hand off to `numerical-optimization-formulation`\n  - then to `cuopt-numerical-optimization-api-python` (or\n    `cuopt-numerical-optimization-api-cli` for MPS inputs)\n\n- If the data suggests QP:\n  - hand off to `numerical-optimization-formulation`\n  - then to `cuopt-numerical-optimization-api-python`\n\n- If the data suggests routing:\n  - hand off to `routing-formulation`\n  - then to `cuopt-routing-api-python`\n\n- If mode selection is still needed because replayability, audit, export, or reuse may matter:\n  - use `optimization-mode-router` before deep model construction\n\n- If optimization intent itself is still uncertain:\n  - use `optimization-intent-router`\n\n## Success criterion\n\nThis skill succeeds when the downstream model-building step can proceed with either:\n- no clarification, or\n- one or a few narrowly targeted clarifying questions\n\nIt fails when it produces a vague restatement of the table without narrowing the modeling interpretation.\n",{"data":44,"body":53},{"name":4,"version":45,"description":6,"license":26,"metadata":46,"origin":52},"26.06.01",{"author":47,"tags":48},"NVIDIA cuOpt Team",[49,50,51],"cuopt","nemoclaw","orchestration","skill-evolution",{"type":54,"children":55},"root",[56,64,70,75,88,107,112,119,124,144,149,182,187,193,198,231,237,244,249,262,267,300,305,311,316,321,364,369,417,422,470,476,481,485,518,523,529,534,541,546,579,585,589,617,623,627,665,671,676,680,708,719,725,730,788,793,799,804,810,915,921,997,1003,1008,1036,1042,1054,1059,1087,1093,1105,1185,1229,1235,1240,1258,1263,1269,1275,1280,1341,1346,1387,1393,1397,1465,1469,1492,1498,1503,1598,1602,1652,1658,1663,1708,1712,1725,1731,1778,1784,1920,1926,1931,1944],{"type":57,"tag":58,"props":59,"children":60},"element","h1",{"id":4},[61],{"type":62,"value":63},"text","Tabular Optimization Ingestion",{"type":57,"tag":65,"props":66,"children":67},"p",{},[68],{"type":62,"value":69},"Use this skill when the user provides raw or semi-structured data and asks a question that may require optimization.",{"type":57,"tag":65,"props":71,"children":72},{},[73],{"type":62,"value":74},"The purpose of this skill is to bridge the gap between messy uploaded data and solver-ready model construction.",{"type":57,"tag":65,"props":76,"children":77},{},[78,80,86],{"type":62,"value":79},"This skill does ",{"type":57,"tag":81,"props":82,"children":83},"strong",{},[84],{"type":62,"value":85},"not",{"type":62,"value":87}," solve the optimization problem itself. It inspects the data, infers likely modeling roles, and identifies what still needs clarification.",{"type":57,"tag":65,"props":89,"children":90},{},[91,96,98,105],{"type":57,"tag":81,"props":92,"children":93},{},[94],{"type":62,"value":95},"It does not authorize heuristic, greedy, or backtracking schedules as\nanswers.",{"type":62,"value":97}," In the NemoClaw sandbox, read ",{"type":57,"tag":99,"props":100,"children":102},"code",{"className":101},[],[103],{"type":62,"value":104},"cuopt-sandbox",{"type":62,"value":106},": the first solver\nthat produces assignments or a schedule must be cuOpt after probe → env →\nsmoke gates pass. Ingestion output is a modeling interpretation (entities,\nobjective fields, constraints) — never a completed plan.",{"type":57,"tag":65,"props":108,"children":109},{},[110],{"type":62,"value":111},"This skill refines the optimization interpretation using the uploaded data; it does not replace the earlier intent decision unless the data clearly contradicts it.",{"type":57,"tag":113,"props":114,"children":116},"h2",{"id":115},"purpose",[117],{"type":62,"value":118},"Purpose",{"type":57,"tag":65,"props":120,"children":121},{},[122],{"type":62,"value":123},"Users do not upload:",{"type":57,"tag":125,"props":126,"children":127},"ul",{},[128,134,139],{"type":57,"tag":129,"props":130,"children":131},"li",{},[132],{"type":62,"value":133},"objective vectors",{"type":57,"tag":129,"props":135,"children":136},{},[137],{"type":62,"value":138},"sparse matrices",{"type":57,"tag":129,"props":140,"children":141},{},[142],{"type":62,"value":143},"explicit decision variable definitions",{"type":57,"tag":65,"props":145,"children":146},{},[147],{"type":62,"value":148},"They upload things like:",{"type":57,"tag":125,"props":150,"children":151},{},[152,157,162,167,172,177],{"type":57,"tag":129,"props":153,"children":154},{},[155],{"type":62,"value":156},"products.csv",{"type":57,"tag":129,"props":158,"children":159},{},[160],{"type":62,"value":161},"capacity.xlsx",{"type":57,"tag":129,"props":163,"children":164},{},[165],{"type":62,"value":166},"orders.json",{"type":57,"tag":129,"props":168,"children":169},{},[170],{"type":62,"value":171},"travel_times.csv",{"type":57,"tag":129,"props":173,"children":174},{},[175],{"type":62,"value":176},"dealers.csv",{"type":57,"tag":129,"props":178,"children":179},{},[180],{"type":62,"value":181},"depots.csv",{"type":57,"tag":65,"props":183,"children":184},{},[185],{"type":62,"value":186},"This skill turns raw tables into a candidate optimization interpretation.",{"type":57,"tag":113,"props":188,"children":190},{"id":189},"what-this-skill-should-produce",[191],{"type":62,"value":192},"What this skill should produce",{"type":57,"tag":65,"props":194,"children":195},{},[196],{"type":62,"value":197},"After inspection, produce a compact working interpretation containing:",{"type":57,"tag":125,"props":199,"children":200},{},[201,206,211,216,221,226],{"type":57,"tag":129,"props":202,"children":203},{},[204],{"type":62,"value":205},"likely problem type: LP \u002F MILP \u002F QP \u002F routing \u002F unclear",{"type":57,"tag":129,"props":207,"children":208},{},[209],{"type":62,"value":210},"likely entities: products, resources, orders, locations, vehicles, customers, depots, time periods, etc.",{"type":57,"tag":129,"props":212,"children":213},{},[214],{"type":62,"value":215},"likely objective fields: profit, cost, margin, time, risk, distance, penalty",{"type":57,"tag":129,"props":217,"children":218},{},[219],{"type":62,"value":220},"likely constraint fields: capacity, demand, budget, inventory, hours, availability, limits, min\u002Fmax bounds",{"type":57,"tag":129,"props":222,"children":223},{},[224],{"type":62,"value":225},"likely decision grain: per product, per route, per order, per plant, per vehicle-stop, etc.",{"type":57,"tag":129,"props":227,"children":228},{},[229],{"type":62,"value":230},"minimum unresolved ambiguities that require a user clarification",{"type":57,"tag":113,"props":232,"children":234},{"id":233},"ingestion-workflow",[235],{"type":62,"value":236},"Ingestion workflow",{"type":57,"tag":238,"props":239,"children":241},"h3",{"id":240},"_1-identify-table-shape-and-grain",[242],{"type":62,"value":243},"1. Identify table shape and grain",{"type":57,"tag":65,"props":245,"children":246},{},[247],{"type":62,"value":248},"For each uploaded source, determine:",{"type":57,"tag":125,"props":250,"children":251},{},[252,257],{"type":57,"tag":129,"props":253,"children":254},{},[255],{"type":62,"value":256},"what one row appears to represent",{"type":57,"tag":129,"props":258,"children":259},{},[260],{"type":62,"value":261},"whether the table is entity-level, transaction-level, matrix-like, or parameter-like",{"type":57,"tag":65,"props":263,"children":264},{},[265],{"type":62,"value":266},"Examples:",{"type":57,"tag":125,"props":268,"children":269},{},[270,275,280,285,290,295],{"type":57,"tag":129,"props":271,"children":272},{},[273],{"type":62,"value":274},"one row per product",{"type":57,"tag":129,"props":276,"children":277},{},[278],{"type":62,"value":279},"one row per plant",{"type":57,"tag":129,"props":281,"children":282},{},[283],{"type":62,"value":284},"one row per customer order",{"type":57,"tag":129,"props":286,"children":287},{},[288],{"type":62,"value":289},"one row per location pair",{"type":57,"tag":129,"props":291,"children":292},{},[293],{"type":62,"value":294},"one row per vehicle",{"type":57,"tag":129,"props":296,"children":297},{},[298],{"type":62,"value":299},"one row per time period",{"type":57,"tag":65,"props":301,"children":302},{},[303],{"type":62,"value":304},"Always state the candidate row meaning before modeling from it.",{"type":57,"tag":238,"props":306,"children":308},{"id":307},"_2-identify-likely-optimization-roles-by-column-meaning",[309],{"type":62,"value":310},"2. Identify likely optimization roles by column meaning",{"type":57,"tag":65,"props":312,"children":313},{},[314],{"type":62,"value":315},"Inspect column names and values for roles such as:",{"type":57,"tag":65,"props":317,"children":318},{},[319],{"type":62,"value":320},"Objective-like fields:",{"type":57,"tag":125,"props":322,"children":323},{},[324,329,334,339,344,349,354,359],{"type":57,"tag":129,"props":325,"children":326},{},[327],{"type":62,"value":328},"profit",{"type":57,"tag":129,"props":330,"children":331},{},[332],{"type":62,"value":333},"revenue",{"type":57,"tag":129,"props":335,"children":336},{},[337],{"type":62,"value":338},"margin",{"type":57,"tag":129,"props":340,"children":341},{},[342],{"type":62,"value":343},"cost",{"type":57,"tag":129,"props":345,"children":346},{},[347],{"type":62,"value":348},"travel_time",{"type":57,"tag":129,"props":350,"children":351},{},[352],{"type":62,"value":353},"distance",{"type":57,"tag":129,"props":355,"children":356},{},[357],{"type":62,"value":358},"risk",{"type":57,"tag":129,"props":360,"children":361},{},[362],{"type":62,"value":363},"penalty",{"type":57,"tag":65,"props":365,"children":366},{},[367],{"type":62,"value":368},"Constraint-like fields:",{"type":57,"tag":125,"props":370,"children":371},{},[372,377,382,387,392,397,402,407,412],{"type":57,"tag":129,"props":373,"children":374},{},[375],{"type":62,"value":376},"capacity",{"type":57,"tag":129,"props":378,"children":379},{},[380],{"type":62,"value":381},"demand",{"type":57,"tag":129,"props":383,"children":384},{},[385],{"type":62,"value":386},"available_hours",{"type":57,"tag":129,"props":388,"children":389},{},[390],{"type":62,"value":391},"budget",{"type":57,"tag":129,"props":393,"children":394},{},[395],{"type":62,"value":396},"max_load",{"type":57,"tag":129,"props":398,"children":399},{},[400],{"type":62,"value":401},"min_order",{"type":57,"tag":129,"props":403,"children":404},{},[405],{"type":62,"value":406},"inventory",{"type":57,"tag":129,"props":408,"children":409},{},[410],{"type":62,"value":411},"time_window_start",{"type":57,"tag":129,"props":413,"children":414},{},[415],{"type":62,"value":416},"time_window_end",{"type":57,"tag":65,"props":418,"children":419},{},[420],{"type":62,"value":421},"Identifier \u002F relationship fields:",{"type":57,"tag":125,"props":423,"children":424},{},[425,430,435,440,445,450,455,460,465],{"type":57,"tag":129,"props":426,"children":427},{},[428],{"type":62,"value":429},"product_id",{"type":57,"tag":129,"props":431,"children":432},{},[433],{"type":62,"value":434},"plant_id",{"type":57,"tag":129,"props":436,"children":437},{},[438],{"type":62,"value":439},"order_id",{"type":57,"tag":129,"props":441,"children":442},{},[443],{"type":62,"value":444},"customer_id",{"type":57,"tag":129,"props":446,"children":447},{},[448],{"type":62,"value":449},"vehicle_id",{"type":57,"tag":129,"props":451,"children":452},{},[453],{"type":62,"value":454},"depot_id",{"type":57,"tag":129,"props":456,"children":457},{},[458],{"type":62,"value":459},"location_id",{"type":57,"tag":129,"props":461,"children":462},{},[463],{"type":62,"value":464},"origin \u002F destination",{"type":57,"tag":129,"props":466,"children":467},{},[468],{"type":62,"value":469},"pickup \u002F delivery",{"type":57,"tag":238,"props":471,"children":473},{"id":472},"_3-infer-candidate-decision-structure",[474],{"type":62,"value":475},"3. Infer candidate decision structure",{"type":57,"tag":65,"props":477,"children":478},{},[479],{"type":62,"value":480},"Infer what the model is probably deciding.",{"type":57,"tag":65,"props":482,"children":483},{},[484],{"type":62,"value":266},{"type":57,"tag":125,"props":486,"children":487},{},[488,493,498,503,508,513],{"type":57,"tag":129,"props":489,"children":490},{},[491],{"type":62,"value":492},"units to produce by product",{"type":57,"tag":129,"props":494,"children":495},{},[496],{"type":62,"value":497},"units to ship from origin to destination",{"type":57,"tag":129,"props":499,"children":500},{},[501],{"type":62,"value":502},"whether to open a facility",{"type":57,"tag":129,"props":504,"children":505},{},[506],{"type":62,"value":507},"which vehicle serves which stops",{"type":57,"tag":129,"props":509,"children":510},{},[511],{"type":62,"value":512},"route sequence through locations",{"type":57,"tag":129,"props":514,"children":515},{},[516],{"type":62,"value":517},"allocation of budget across campaigns",{"type":57,"tag":65,"props":519,"children":520},{},[521],{"type":62,"value":522},"Do not overcommit when the data supports multiple plausible decisions. Note the candidates and ask one focused question if needed.",{"type":57,"tag":238,"props":524,"children":526},{"id":525},"_4-match-the-data-pattern-to-a-problem-family",[527],{"type":62,"value":528},"4. Match the data pattern to a problem family",{"type":57,"tag":65,"props":530,"children":531},{},[532],{"type":62,"value":533},"Use common table patterns:",{"type":57,"tag":535,"props":536,"children":538},"h4",{"id":537},"lp-milp-patterns",[539],{"type":62,"value":540},"LP \u002F MILP patterns",{"type":57,"tag":65,"props":542,"children":543},{},[544],{"type":62,"value":545},"Typical signs:",{"type":57,"tag":125,"props":547,"children":548},{},[549,554,559,564,569,574],{"type":57,"tag":129,"props":550,"children":551},{},[552],{"type":62,"value":553},"product\u002Fresource tables",{"type":57,"tag":129,"props":555,"children":556},{},[557],{"type":62,"value":558},"profit and resource consumption by product",{"type":57,"tag":129,"props":560,"children":561},{},[562],{"type":62,"value":563},"capacity tables",{"type":57,"tag":129,"props":565,"children":566},{},[567],{"type":62,"value":568},"demand tables",{"type":57,"tag":129,"props":570,"children":571},{},[572],{"type":62,"value":573},"assignment tables",{"type":57,"tag":129,"props":575,"children":576},{},[577],{"type":62,"value":578},"open\u002Fclose or yes\u002Fno columns implying integrality",{"type":57,"tag":535,"props":580,"children":582},{"id":581},"qp-patterns",[583],{"type":62,"value":584},"QP patterns",{"type":57,"tag":65,"props":586,"children":587},{},[588],{"type":62,"value":545},{"type":57,"tag":125,"props":590,"children":591},{},[592,597,602,607,612],{"type":57,"tag":129,"props":593,"children":594},{},[595],{"type":62,"value":596},"covariance matrix",{"type":57,"tag":129,"props":598,"children":599},{},[600],{"type":62,"value":601},"risk matrix",{"type":57,"tag":129,"props":603,"children":604},{},[605],{"type":62,"value":606},"variance terms",{"type":57,"tag":129,"props":608,"children":609},{},[610],{"type":62,"value":611},"squared error context",{"type":57,"tag":129,"props":613,"children":614},{},[615],{"type":62,"value":616},"portfolio weights or regression-style structure",{"type":57,"tag":535,"props":618,"children":620},{"id":619},"routing-patterns",[621],{"type":62,"value":622},"Routing patterns",{"type":57,"tag":65,"props":624,"children":625},{},[626],{"type":62,"value":545},{"type":57,"tag":125,"props":628,"children":629},{},[630,635,640,645,650,655,660],{"type":57,"tag":129,"props":631,"children":632},{},[633],{"type":62,"value":634},"locations \u002F coordinates",{"type":57,"tag":129,"props":636,"children":637},{},[638],{"type":62,"value":639},"travel time or distance matrix",{"type":57,"tag":129,"props":641,"children":642},{},[643],{"type":62,"value":644},"depot + customer tables",{"type":57,"tag":129,"props":646,"children":647},{},[648],{"type":62,"value":649},"vehicle tables",{"type":57,"tag":129,"props":651,"children":652},{},[653],{"type":62,"value":654},"demands and capacities",{"type":57,"tag":129,"props":656,"children":657},{},[658],{"type":62,"value":659},"time windows",{"type":57,"tag":129,"props":661,"children":662},{},[663],{"type":62,"value":664},"pickup and delivery pairs",{"type":57,"tag":238,"props":666,"children":668},{"id":667},"_5-identify-only-the-critical-ambiguities",[669],{"type":62,"value":670},"5. Identify only the critical ambiguities",{"type":57,"tag":65,"props":672,"children":673},{},[674],{"type":62,"value":675},"Before handing off, identify the smallest set of unanswered questions that block valid model construction.",{"type":57,"tag":65,"props":677,"children":678},{},[679],{"type":62,"value":266},{"type":57,"tag":125,"props":681,"children":682},{},[683,688,693,698,703],{"type":57,"tag":129,"props":684,"children":685},{},[686],{"type":62,"value":687},"Are demands mandatory or forecast-only?",{"type":57,"tag":129,"props":689,"children":690},{},[691],{"type":62,"value":692},"Must decisions be integers?",{"type":57,"tag":129,"props":694,"children":695},{},[696],{"type":62,"value":697},"Are these time windows hard constraints?",{"type":57,"tag":129,"props":699,"children":700},{},[701],{"type":62,"value":702},"Can unmet demand be allowed with penalty?",{"type":57,"tag":129,"props":704,"children":705},{},[706],{"type":62,"value":707},"Is profit net profit or revenue only?",{"type":57,"tag":65,"props":709,"children":710},{},[711,713,717],{"type":62,"value":712},"Do ",{"type":57,"tag":81,"props":714,"children":715},{},[716],{"type":62,"value":85},{"type":62,"value":718}," ask broad generic questions if the data already strongly suggests the answer.",{"type":57,"tag":113,"props":720,"children":722},{"id":721},"preferred-output-format",[723],{"type":62,"value":724},"Preferred output format",{"type":57,"tag":65,"props":726,"children":727},{},[728],{"type":62,"value":729},"Summarize findings compactly:",{"type":57,"tag":125,"props":731,"children":732},{},[733,743,752,761,770,779],{"type":57,"tag":129,"props":734,"children":735},{},[736,741],{"type":57,"tag":81,"props":737,"children":738},{},[739],{"type":62,"value":740},"Candidate problem family:",{"type":62,"value":742}," ...",{"type":57,"tag":129,"props":744,"children":745},{},[746,751],{"type":57,"tag":81,"props":747,"children":748},{},[749],{"type":62,"value":750},"Row meanings:",{"type":62,"value":742},{"type":57,"tag":129,"props":753,"children":754},{},[755,760],{"type":57,"tag":81,"props":756,"children":757},{},[758],{"type":62,"value":759},"Likely decision:",{"type":62,"value":742},{"type":57,"tag":129,"props":762,"children":763},{},[764,769],{"type":57,"tag":81,"props":765,"children":766},{},[767],{"type":62,"value":768},"Likely objective:",{"type":62,"value":742},{"type":57,"tag":129,"props":771,"children":772},{},[773,778],{"type":57,"tag":81,"props":774,"children":775},{},[776],{"type":62,"value":777},"Likely constraints:",{"type":62,"value":742},{"type":57,"tag":129,"props":780,"children":781},{},[782,787],{"type":57,"tag":81,"props":783,"children":784},{},[785],{"type":62,"value":786},"Unresolved blockers:",{"type":62,"value":742},{"type":57,"tag":65,"props":789,"children":790},{},[791],{"type":62,"value":792},"This summary should be short enough that a downstream formulation skill can use it directly.",{"type":57,"tag":113,"props":794,"children":796},{"id":795},"column-role-heuristics",[797],{"type":62,"value":798},"Column-role heuristics",{"type":57,"tag":65,"props":800,"children":801},{},[802],{"type":62,"value":803},"Use these heuristics carefully. They guide interpretation but do not prove it.",{"type":57,"tag":238,"props":805,"children":807},{"id":806},"cost-objective-heuristics",[808],{"type":62,"value":809},"Cost \u002F objective heuristics",{"type":57,"tag":125,"props":811,"children":812},{},[813,845,868,891],{"type":57,"tag":129,"props":814,"children":815},{},[816,821,823,829,830,836,837,843],{"type":57,"tag":99,"props":817,"children":819},{"className":818},[],[820],{"type":62,"value":343},{"type":62,"value":822},", ",{"type":57,"tag":99,"props":824,"children":826},{"className":825},[],[827],{"type":62,"value":828},"unit_cost",{"type":62,"value":822},{"type":57,"tag":99,"props":831,"children":833},{"className":832},[],[834],{"type":62,"value":835},"shipping_cost",{"type":62,"value":822},{"type":57,"tag":99,"props":838,"children":840},{"className":839},[],[841],{"type":62,"value":842},"expense",{"type":62,"value":844}," → likely minimization coefficient",{"type":57,"tag":129,"props":846,"children":847},{},[848,853,854,859,860,866],{"type":57,"tag":99,"props":849,"children":851},{"className":850},[],[852],{"type":62,"value":328},{"type":62,"value":822},{"type":57,"tag":99,"props":855,"children":857},{"className":856},[],[858],{"type":62,"value":338},{"type":62,"value":822},{"type":57,"tag":99,"props":861,"children":863},{"className":862},[],[864],{"type":62,"value":865},"contribution",{"type":62,"value":867}," → likely maximization coefficient",{"type":57,"tag":129,"props":869,"children":870},{},[871,876,877,882,883,889],{"type":57,"tag":99,"props":872,"children":874},{"className":873},[],[875],{"type":62,"value":353},{"type":62,"value":822},{"type":57,"tag":99,"props":878,"children":880},{"className":879},[],[881],{"type":62,"value":348},{"type":62,"value":822},{"type":57,"tag":99,"props":884,"children":886},{"className":885},[],[887],{"type":62,"value":888},"duration",{"type":62,"value":890}," → likely routing objective term or service constraint",{"type":57,"tag":129,"props":892,"children":893},{},[894,899,900,906,907,913],{"type":57,"tag":99,"props":895,"children":897},{"className":896},[],[898],{"type":62,"value":358},{"type":62,"value":822},{"type":57,"tag":99,"props":901,"children":903},{"className":902},[],[904],{"type":62,"value":905},"variance",{"type":62,"value":822},{"type":57,"tag":99,"props":908,"children":910},{"className":909},[],[911],{"type":62,"value":912},"covariance",{"type":62,"value":914}," → likely QP signal",{"type":57,"tag":238,"props":916,"children":918},{"id":917},"capacity-demand-heuristics",[919],{"type":62,"value":920},"Capacity \u002F demand heuristics",{"type":57,"tag":125,"props":922,"children":923},{},[924,955,979],{"type":57,"tag":129,"props":925,"children":926},{},[927,932,933,939,940,946,947,953],{"type":57,"tag":99,"props":928,"children":930},{"className":929},[],[931],{"type":62,"value":376},{"type":62,"value":822},{"type":57,"tag":99,"props":934,"children":936},{"className":935},[],[937],{"type":62,"value":938},"available",{"type":62,"value":822},{"type":57,"tag":99,"props":941,"children":943},{"className":942},[],[944],{"type":62,"value":945},"limit",{"type":62,"value":822},{"type":57,"tag":99,"props":948,"children":950},{"className":949},[],[951],{"type":62,"value":952},"max_*",{"type":62,"value":954}," → upper bound \u002F resource constraint",{"type":57,"tag":129,"props":956,"children":957},{},[958,963,964,970,971,977],{"type":57,"tag":99,"props":959,"children":961},{"className":960},[],[962],{"type":62,"value":381},{"type":62,"value":822},{"type":57,"tag":99,"props":965,"children":967},{"className":966},[],[968],{"type":62,"value":969},"required",{"type":62,"value":822},{"type":57,"tag":99,"props":972,"children":974},{"className":973},[],[975],{"type":62,"value":976},"need",{"type":62,"value":978}," → demand fulfillment or service requirement",{"type":57,"tag":129,"props":980,"children":981},{},[982,988,989,995],{"type":57,"tag":99,"props":983,"children":985},{"className":984},[],[986],{"type":62,"value":987},"min_*",{"type":62,"value":822},{"type":57,"tag":99,"props":990,"children":992},{"className":991},[],[993],{"type":62,"value":994},"minimum_*",{"type":62,"value":996}," → lower bound or service rule",{"type":57,"tag":238,"props":998,"children":1000},{"id":999},"integrality-heuristics",[1001],{"type":62,"value":1002},"Integrality heuristics",{"type":57,"tag":65,"props":1004,"children":1005},{},[1006],{"type":62,"value":1007},"Treat MILP as likely when the data or request suggests:",{"type":57,"tag":125,"props":1009,"children":1010},{},[1011,1016,1021,1026,1031],{"type":57,"tag":129,"props":1012,"children":1013},{},[1014],{"type":62,"value":1015},"whole units",{"type":57,"tag":129,"props":1017,"children":1018},{},[1019],{"type":62,"value":1020},"counts of vehicles, workers, facilities, shifts, items",{"type":57,"tag":129,"props":1022,"children":1023},{},[1024],{"type":62,"value":1025},"yes\u002Fno choices",{"type":57,"tag":129,"props":1027,"children":1028},{},[1029],{"type":62,"value":1030},"on\u002Foff decisions",{"type":57,"tag":129,"props":1032,"children":1033},{},[1034],{"type":62,"value":1035},"assignment indicators",{"type":57,"tag":238,"props":1037,"children":1039},{"id":1038},"routing-heuristics",[1040],{"type":62,"value":1041},"Routing heuristics",{"type":57,"tag":65,"props":1043,"children":1044},{},[1045,1047,1052],{"type":62,"value":1046},"Treat routing as likely when the core question depends on ",{"type":57,"tag":81,"props":1048,"children":1049},{},[1050],{"type":62,"value":1051},"path construction",{"type":62,"value":1053},", not merely allocation.",{"type":57,"tag":65,"props":1055,"children":1056},{},[1057],{"type":62,"value":1058},"Signs include:",{"type":57,"tag":125,"props":1060,"children":1061},{},[1062,1067,1072,1077,1082],{"type":57,"tag":129,"props":1063,"children":1064},{},[1065],{"type":62,"value":1066},"stop sequence matters",{"type":57,"tag":129,"props":1068,"children":1069},{},[1070],{"type":62,"value":1071},"travel between locations matters",{"type":57,"tag":129,"props":1073,"children":1074},{},[1075],{"type":62,"value":1076},"vehicles start\u002Fend at depots",{"type":57,"tag":129,"props":1078,"children":1079},{},[1080],{"type":62,"value":1081},"travel matrix or coordinates are present",{"type":57,"tag":129,"props":1083,"children":1084},{},[1085],{"type":62,"value":1086},"pickup and delivery relationships appear",{"type":57,"tag":238,"props":1088,"children":1090},{"id":1089},"cross-row-coupling-heuristics",[1091],{"type":62,"value":1092},"Cross-row coupling heuristics",{"type":57,"tag":65,"props":1094,"children":1095},{},[1096,1098,1103],{"type":62,"value":1097},"Duplicate values in a foreign-key column across rows of a parent table usually signal a ",{"type":57,"tag":81,"props":1099,"children":1100},{},[1101],{"type":62,"value":1102},"shared agent or resource",{"type":62,"value":1104}," — one entity serving multiple parents. Shared resources need a mutual-exclusion constraint that no other column states.",{"type":57,"tag":125,"props":1106,"children":1107},{},[1108,1147,1172],{"type":57,"tag":129,"props":1109,"children":1110},{},[1111,1117,1118,1124,1125,1131,1132,1138,1139,1145],{"type":57,"tag":99,"props":1112,"children":1114},{"className":1113},[],[1115],{"type":62,"value":1116},"coach_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1119,"children":1121},{"className":1120},[],[1122],{"type":62,"value":1123},"driver_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1126,"children":1128},{"className":1127},[],[1129],{"type":62,"value":1130},"instructor_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1133,"children":1135},{"className":1134},[],[1136],{"type":62,"value":1137},"nurse_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1140,"children":1142},{"className":1141},[],[1143],{"type":62,"value":1144},"operator_id",{"type":62,"value":1146}," → shared agent; can serve only one parent at a time",{"type":57,"tag":129,"props":1148,"children":1149},{},[1150,1156,1157,1163,1164,1170],{"type":57,"tag":99,"props":1151,"children":1153},{"className":1152},[],[1154],{"type":62,"value":1155},"machine_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1158,"children":1160},{"className":1159},[],[1161],{"type":62,"value":1162},"bay_id",{"type":62,"value":822},{"type":57,"tag":99,"props":1165,"children":1167},{"className":1166},[],[1168],{"type":62,"value":1169},"tool_id",{"type":62,"value":1171}," → shared resource; can host only one job at a time",{"type":57,"tag":129,"props":1173,"children":1174},{},[1175,1177,1183],{"type":62,"value":1176},"any FK column where ",{"type":57,"tag":99,"props":1178,"children":1180},{"className":1179},[],[1181],{"type":62,"value":1182},"distinct_values \u003C row_count",{"type":62,"value":1184}," → check whether simultaneous assignment is allowed",{"type":57,"tag":65,"props":1186,"children":1187},{},[1188,1190,1196,1198,1204,1206,1212,1214,1219,1221,1227],{"type":62,"value":1189},"A ",{"type":57,"tag":99,"props":1191,"children":1193},{"className":1192},[],[1194],{"type":62,"value":1195},"*_unavailability",{"type":62,"value":1197}," (or ",{"type":57,"tag":99,"props":1199,"children":1201},{"className":1200},[],[1202],{"type":62,"value":1203},"*_availability",{"type":62,"value":1205},") table documents ",{"type":57,"tag":1207,"props":1208,"children":1209},"em",{},[1210],{"type":62,"value":1211},"known absences",{"type":62,"value":1213},"; duplicated FK values document ",{"type":57,"tag":1207,"props":1215,"children":1216},{},[1217],{"type":62,"value":1218},"implicit conflicts",{"type":62,"value":1220},". Treat both as constraint sources. Concrete check: for every FK-looking column in a parent table, compare distinct value count to row count, and surface the column when ",{"type":57,"tag":99,"props":1222,"children":1224},{"className":1223},[],[1225],{"type":62,"value":1226},"distinct \u003C rows",{"type":62,"value":1228},".",{"type":57,"tag":238,"props":1230,"children":1232},{"id":1231},"pre-solve-capacity-arithmetic",[1233],{"type":62,"value":1234},"Pre-solve capacity arithmetic",{"type":57,"tag":65,"props":1236,"children":1237},{},[1238],{"type":62,"value":1239},"Before handing off to formulation, compare aggregate requirements against\nupper bounds implied by the tables:",{"type":57,"tag":125,"props":1241,"children":1242},{},[1243,1248,1253],{"type":57,"tag":129,"props":1244,"children":1245},{},[1246],{"type":62,"value":1247},"required events or appearances per entity (e.g. round-robin games per team)",{"type":57,"tag":129,"props":1249,"children":1250},{},[1251],{"type":62,"value":1252},"maximum placements allowed by per-week, per-day, or per-slot caps",{"type":57,"tag":129,"props":1254,"children":1255},{},[1256],{"type":62,"value":1257},"total court\u002Fslot\u002Fmachine capacity across the horizon",{"type":57,"tag":65,"props":1259,"children":1260},{},[1261],{"type":62,"value":1262},"If required appearances exceed what the time structure can host, flag the\nconflict before building the full MILP. Example: 8 required games per team\nwith at most 1 game per team per week over 5 weeks allows only 5 games —\ninfeasible without relaxing the weekly cap.",{"type":57,"tag":113,"props":1264,"children":1266},{"id":1265},"examples",[1267],{"type":62,"value":1268},"Examples",{"type":57,"tag":238,"props":1270,"children":1272},{"id":1271},"example-1-product-mix-tables",[1273],{"type":62,"value":1274},"Example 1: product mix tables",{"type":57,"tag":65,"props":1276,"children":1277},{},[1278],{"type":62,"value":1279},"Files include:",{"type":57,"tag":125,"props":1281,"children":1282},{},[1283,1319],{"type":57,"tag":129,"props":1284,"children":1285},{},[1286,1291,1293,1299,1300,1305,1306,1312,1313],{"type":57,"tag":99,"props":1287,"children":1289},{"className":1288},[],[1290],{"type":62,"value":156},{"type":62,"value":1292}," with columns like ",{"type":57,"tag":99,"props":1294,"children":1296},{"className":1295},[],[1297],{"type":62,"value":1298},"product",{"type":62,"value":822},{"type":57,"tag":99,"props":1301,"children":1303},{"className":1302},[],[1304],{"type":62,"value":328},{"type":62,"value":822},{"type":57,"tag":99,"props":1307,"children":1309},{"className":1308},[],[1310],{"type":62,"value":1311},"labor_hours",{"type":62,"value":822},{"type":57,"tag":99,"props":1314,"children":1316},{"className":1315},[],[1317],{"type":62,"value":1318},"steel_units",{"type":57,"tag":129,"props":1320,"children":1321},{},[1322,1328,1329,1335,1336],{"type":57,"tag":99,"props":1323,"children":1325},{"className":1324},[],[1326],{"type":62,"value":1327},"capacity.csv",{"type":62,"value":1292},{"type":57,"tag":99,"props":1330,"children":1332},{"className":1331},[],[1333],{"type":62,"value":1334},"resource",{"type":62,"value":822},{"type":57,"tag":99,"props":1337,"children":1339},{"className":1338},[],[1340],{"type":62,"value":938},{"type":57,"tag":65,"props":1342,"children":1343},{},[1344],{"type":62,"value":1345},"Likely interpretation:",{"type":57,"tag":125,"props":1347,"children":1348},{},[1349,1361,1372,1377,1382],{"type":57,"tag":129,"props":1350,"children":1351},{},[1352,1354,1359],{"type":62,"value":1353},"one row in ",{"type":57,"tag":99,"props":1355,"children":1357},{"className":1356},[],[1358],{"type":62,"value":156},{"type":62,"value":1360}," = one product",{"type":57,"tag":129,"props":1362,"children":1363},{},[1364,1365,1370],{"type":62,"value":1353},{"type":57,"tag":99,"props":1366,"children":1368},{"className":1367},[],[1369],{"type":62,"value":1327},{"type":62,"value":1371}," = one resource limit",{"type":57,"tag":129,"props":1373,"children":1374},{},[1375],{"type":62,"value":1376},"decision = how much of each product to produce",{"type":57,"tag":129,"props":1378,"children":1379},{},[1380],{"type":62,"value":1381},"objective = maximize profit",{"type":57,"tag":129,"props":1383,"children":1384},{},[1385],{"type":62,"value":1386},"constraints = labor and steel capacities",{"type":57,"tag":238,"props":1388,"children":1390},{"id":1389},"example-2-routing-tables",[1391],{"type":62,"value":1392},"Example 2: routing tables",{"type":57,"tag":65,"props":1394,"children":1395},{},[1396],{"type":62,"value":1279},{"type":57,"tag":125,"props":1398,"children":1399},{},[1400,1434,1455],{"type":57,"tag":129,"props":1401,"children":1402},{},[1403,1409,1411,1416,1417,1422,1423,1428,1429],{"type":57,"tag":99,"props":1404,"children":1406},{"className":1405},[],[1407],{"type":62,"value":1408},"customers.csv",{"type":62,"value":1410}," with ",{"type":57,"tag":99,"props":1412,"children":1414},{"className":1413},[],[1415],{"type":62,"value":444},{"type":62,"value":822},{"type":57,"tag":99,"props":1418,"children":1420},{"className":1419},[],[1421],{"type":62,"value":381},{"type":62,"value":822},{"type":57,"tag":99,"props":1424,"children":1426},{"className":1425},[],[1427],{"type":62,"value":411},{"type":62,"value":822},{"type":57,"tag":99,"props":1430,"children":1432},{"className":1431},[],[1433],{"type":62,"value":416},{"type":57,"tag":129,"props":1435,"children":1436},{},[1437,1443,1444,1449,1450],{"type":57,"tag":99,"props":1438,"children":1440},{"className":1439},[],[1441],{"type":62,"value":1442},"vehicles.csv",{"type":62,"value":1410},{"type":57,"tag":99,"props":1445,"children":1447},{"className":1446},[],[1448],{"type":62,"value":449},{"type":62,"value":822},{"type":57,"tag":99,"props":1451,"children":1453},{"className":1452},[],[1454],{"type":62,"value":376},{"type":57,"tag":129,"props":1456,"children":1457},{},[1458,1463],{"type":57,"tag":99,"props":1459,"children":1461},{"className":1460},[],[1462],{"type":62,"value":171},{"type":62,"value":1464}," with origin\u002Fdestination or matrix-style travel times",{"type":57,"tag":65,"props":1466,"children":1467},{},[1468],{"type":62,"value":1345},{"type":57,"tag":125,"props":1470,"children":1471},{},[1472,1477,1482,1487],{"type":57,"tag":129,"props":1473,"children":1474},{},[1475],{"type":62,"value":1476},"customer rows are stops with demand and optional time windows",{"type":57,"tag":129,"props":1478,"children":1479},{},[1480],{"type":62,"value":1481},"vehicle rows define fleet capacity",{"type":57,"tag":129,"props":1483,"children":1484},{},[1485],{"type":62,"value":1486},"travel table defines movement cost\u002Ftime",{"type":57,"tag":129,"props":1488,"children":1489},{},[1490],{"type":62,"value":1491},"likely problem family = routing",{"type":57,"tag":238,"props":1493,"children":1495},{"id":1494},"example-3-time-slot-resource-assignment-scheduling-milp",[1496],{"type":62,"value":1497},"Example 3: time-slot \u002F resource assignment (scheduling MILP)",{"type":57,"tag":65,"props":1499,"children":1500},{},[1501],{"type":62,"value":1502},"Files include patterns such as:",{"type":57,"tag":125,"props":1504,"children":1505},{},[1506,1525,1543,1569,1587],{"type":57,"tag":129,"props":1507,"children":1508},{},[1509,1515,1517,1523],{"type":57,"tag":99,"props":1510,"children":1512},{"className":1511},[],[1513],{"type":62,"value":1514},"games.csv",{"type":62,"value":1516}," or ",{"type":57,"tag":99,"props":1518,"children":1520},{"className":1519},[],[1521],{"type":62,"value":1522},"jobs.csv",{"type":62,"value":1524}," — items to place (events, tasks, orders)",{"type":57,"tag":129,"props":1526,"children":1527},{},[1528,1534,1535,1541],{"type":57,"tag":99,"props":1529,"children":1531},{"className":1530},[],[1532],{"type":62,"value":1533},"time_slots.csv",{"type":62,"value":1516},{"type":57,"tag":99,"props":1536,"children":1538},{"className":1537},[],[1539],{"type":62,"value":1540},"shifts.csv",{"type":62,"value":1542}," — when placement can occur",{"type":57,"tag":129,"props":1544,"children":1545},{},[1546,1552,1553,1559,1561,1567],{"type":57,"tag":99,"props":1547,"children":1549},{"className":1548},[],[1550],{"type":62,"value":1551},"courts.csv",{"type":62,"value":822},{"type":57,"tag":99,"props":1554,"children":1556},{"className":1555},[],[1557],{"type":62,"value":1558},"machines.csv",{"type":62,"value":1560},", or ",{"type":57,"tag":99,"props":1562,"children":1564},{"className":1563},[],[1565],{"type":62,"value":1566},"rooms.csv",{"type":62,"value":1568}," — resources",{"type":57,"tag":129,"props":1570,"children":1571},{},[1572,1578,1579,1585],{"type":57,"tag":99,"props":1573,"children":1575},{"className":1574},[],[1576],{"type":62,"value":1577},"teams.csv",{"type":62,"value":1516},{"type":57,"tag":99,"props":1580,"children":1582},{"className":1581},[],[1583],{"type":62,"value":1584},"workers.csv",{"type":62,"value":1586}," — entities tied to shared agents (coaches, operators)",{"type":57,"tag":129,"props":1588,"children":1589},{},[1590,1596],{"type":57,"tag":99,"props":1591,"children":1593},{"className":1592},[],[1594],{"type":62,"value":1595},"*_unavailability.csv",{"type":62,"value":1597}," — blocked (resource, slot) or (agent, slot) pairs",{"type":57,"tag":65,"props":1599,"children":1600},{},[1601],{"type":62,"value":1345},{"type":57,"tag":125,"props":1603,"children":1604},{},[1605,1610,1615,1627],{"type":57,"tag":129,"props":1606,"children":1607},{},[1608],{"type":62,"value":1609},"decision = assign each item to a (slot, resource) or similar binary\u002Finteger placement",{"type":57,"tag":129,"props":1611,"children":1612},{},[1613],{"type":62,"value":1614},"hard constraints = no double-booking, unavailability, capacity, one game per team per slot",{"type":57,"tag":129,"props":1616,"children":1617},{},[1618,1620,1625],{"type":62,"value":1619},"likely problem family = ",{"type":57,"tag":81,"props":1621,"children":1622},{},[1623],{"type":62,"value":1624},"MILP",{"type":62,"value":1626}," (even if user only says \"build a schedule\" or \"valid plan\")",{"type":57,"tag":129,"props":1628,"children":1629},{},[1630,1635,1637,1643,1645,1650],{"type":57,"tag":81,"props":1631,"children":1632},{},[1633],{"type":62,"value":1634},"NemoClaw:",{"type":62,"value":1636}," read ",{"type":57,"tag":99,"props":1638,"children":1640},{"className":1639},[],[1641],{"type":62,"value":1642},"optimization-from-data-orchestrator",{"type":62,"value":1644}," + ",{"type":57,"tag":99,"props":1646,"children":1648},{"className":1647},[],[1649],{"type":62,"value":104},{"type":62,"value":1651}," before any custom scheduler code",{"type":57,"tag":238,"props":1653,"children":1655},{"id":1654},"example-4-historical-transaction-table",[1656],{"type":62,"value":1657},"Example 4: historical transaction table",{"type":57,"tag":65,"props":1659,"children":1660},{},[1661],{"type":62,"value":1662},"File includes:",{"type":57,"tag":125,"props":1664,"children":1665},{},[1666],{"type":57,"tag":129,"props":1667,"children":1668},{},[1669,1675,1676,1681,1682,1688,1689,1695,1696,1701,1702],{"type":57,"tag":99,"props":1670,"children":1672},{"className":1671},[],[1673],{"type":62,"value":1674},"sales_history.csv",{"type":62,"value":1410},{"type":57,"tag":99,"props":1677,"children":1679},{"className":1678},[],[1680],{"type":62,"value":439},{"type":62,"value":822},{"type":57,"tag":99,"props":1683,"children":1685},{"className":1684},[],[1686],{"type":62,"value":1687},"date",{"type":62,"value":822},{"type":57,"tag":99,"props":1690,"children":1692},{"className":1691},[],[1693],{"type":62,"value":1694},"region",{"type":62,"value":822},{"type":57,"tag":99,"props":1697,"children":1699},{"className":1698},[],[1700],{"type":62,"value":333},{"type":62,"value":822},{"type":57,"tag":99,"props":1703,"children":1705},{"className":1704},[],[1706],{"type":62,"value":1707},"units_sold",{"type":57,"tag":65,"props":1709,"children":1710},{},[1711],{"type":62,"value":1345},{"type":57,"tag":125,"props":1713,"children":1714},{},[1715,1720],{"type":57,"tag":129,"props":1716,"children":1717},{},[1718],{"type":62,"value":1719},"this may be analytics, not optimization, unless the user asks for a future decision under constraints",{"type":57,"tag":129,"props":1721,"children":1722},{},[1723],{"type":62,"value":1724},"do not invent decision variables from history alone without a decision-focused question",{"type":57,"tag":113,"props":1726,"children":1728},{"id":1727},"guardrails",[1729],{"type":62,"value":1730},"Guardrails",{"type":57,"tag":125,"props":1732,"children":1733},{},[1734,1753,1758,1763,1768,1773],{"type":57,"tag":129,"props":1735,"children":1736},{},[1737,1739,1744,1746,1751],{"type":62,"value":1738},"Do not assume every table with ",{"type":57,"tag":99,"props":1740,"children":1742},{"className":1741},[],[1743],{"type":62,"value":343},{"type":62,"value":1745}," and ",{"type":57,"tag":99,"props":1747,"children":1749},{"className":1748},[],[1750],{"type":62,"value":376},{"type":62,"value":1752}," is automatically a valid optimization model.",{"type":57,"tag":129,"props":1754,"children":1755},{},[1756],{"type":62,"value":1757},"Do not mistake transaction history for decision variables without evidence.",{"type":57,"tag":129,"props":1759,"children":1760},{},[1761],{"type":62,"value":1762},"Do not treat forecast data as hard constraints unless the user implies that.",{"type":57,"tag":129,"props":1764,"children":1765},{},[1766],{"type":62,"value":1767},"Do not infer QP unless there is a real quadratic signal.",{"type":57,"tag":129,"props":1769,"children":1770},{},[1771],{"type":62,"value":1772},"Do not infer routing unless movement between locations is central to the decision.",{"type":57,"tag":129,"props":1774,"children":1775},{},[1776],{"type":62,"value":1777},"Prefer one precise clarification over a long list of speculative questions.",{"type":57,"tag":113,"props":1779,"children":1781},{"id":1780},"handoff-guidance",[1782],{"type":62,"value":1783},"Handoff guidance",{"type":57,"tag":125,"props":1785,"children":1786},{},[1787,1827,1853,1881,1902],{"type":57,"tag":129,"props":1788,"children":1789},{},[1790,1792],{"type":62,"value":1791},"If the data suggests LP \u002F MILP:",{"type":57,"tag":125,"props":1793,"children":1794},{},[1795,1806],{"type":57,"tag":129,"props":1796,"children":1797},{},[1798,1800],{"type":62,"value":1799},"hand off to ",{"type":57,"tag":99,"props":1801,"children":1803},{"className":1802},[],[1804],{"type":62,"value":1805},"numerical-optimization-formulation",{"type":57,"tag":129,"props":1807,"children":1808},{},[1809,1811,1817,1819,1825],{"type":62,"value":1810},"then to ",{"type":57,"tag":99,"props":1812,"children":1814},{"className":1813},[],[1815],{"type":62,"value":1816},"cuopt-numerical-optimization-api-python",{"type":62,"value":1818}," (or\n",{"type":57,"tag":99,"props":1820,"children":1822},{"className":1821},[],[1823],{"type":62,"value":1824},"cuopt-numerical-optimization-api-cli",{"type":62,"value":1826}," for MPS inputs)",{"type":57,"tag":129,"props":1828,"children":1829},{},[1830,1832],{"type":62,"value":1831},"If the data suggests QP:",{"type":57,"tag":125,"props":1833,"children":1834},{},[1835,1844],{"type":57,"tag":129,"props":1836,"children":1837},{},[1838,1839],{"type":62,"value":1799},{"type":57,"tag":99,"props":1840,"children":1842},{"className":1841},[],[1843],{"type":62,"value":1805},{"type":57,"tag":129,"props":1845,"children":1846},{},[1847,1848],{"type":62,"value":1810},{"type":57,"tag":99,"props":1849,"children":1851},{"className":1850},[],[1852],{"type":62,"value":1816},{"type":57,"tag":129,"props":1854,"children":1855},{},[1856,1858],{"type":62,"value":1857},"If the data suggests routing:",{"type":57,"tag":125,"props":1859,"children":1860},{},[1861,1871],{"type":57,"tag":129,"props":1862,"children":1863},{},[1864,1865],{"type":62,"value":1799},{"type":57,"tag":99,"props":1866,"children":1868},{"className":1867},[],[1869],{"type":62,"value":1870},"routing-formulation",{"type":57,"tag":129,"props":1872,"children":1873},{},[1874,1875],{"type":62,"value":1810},{"type":57,"tag":99,"props":1876,"children":1878},{"className":1877},[],[1879],{"type":62,"value":1880},"cuopt-routing-api-python",{"type":57,"tag":129,"props":1882,"children":1883},{},[1884,1886],{"type":62,"value":1885},"If mode selection is still needed because replayability, audit, export, or reuse may matter:",{"type":57,"tag":125,"props":1887,"children":1888},{},[1889],{"type":57,"tag":129,"props":1890,"children":1891},{},[1892,1894,1900],{"type":62,"value":1893},"use ",{"type":57,"tag":99,"props":1895,"children":1897},{"className":1896},[],[1898],{"type":62,"value":1899},"optimization-mode-router",{"type":62,"value":1901}," before deep model construction",{"type":57,"tag":129,"props":1903,"children":1904},{},[1905,1907],{"type":62,"value":1906},"If optimization intent itself is still uncertain:",{"type":57,"tag":125,"props":1908,"children":1909},{},[1910],{"type":57,"tag":129,"props":1911,"children":1912},{},[1913,1914],{"type":62,"value":1893},{"type":57,"tag":99,"props":1915,"children":1917},{"className":1916},[],[1918],{"type":62,"value":1919},"optimization-intent-router",{"type":57,"tag":113,"props":1921,"children":1923},{"id":1922},"success-criterion",[1924],{"type":62,"value":1925},"Success criterion",{"type":57,"tag":65,"props":1927,"children":1928},{},[1929],{"type":62,"value":1930},"This skill succeeds when the downstream model-building step can proceed with either:",{"type":57,"tag":125,"props":1932,"children":1933},{},[1934,1939],{"type":57,"tag":129,"props":1935,"children":1936},{},[1937],{"type":62,"value":1938},"no clarification, or",{"type":57,"tag":129,"props":1940,"children":1941},{},[1942],{"type":62,"value":1943},"one or a few narrowly targeted clarifying questions",{"type":57,"tag":65,"props":1945,"children":1946},{},[1947],{"type":62,"value":1948},"It fails when it produces a vague restatement of the table without narrowing the modeling interpretation.",{"items":1950,"total":2038},[1951,1966,1976,1990,2003,2017,2029],{"slug":1952,"name":1952,"fn":1953,"description":1954,"org":1955,"tags":1956,"stars":23,"repoUrl":24,"updatedAt":1965},"cuopt-debugging","debug NVIDIA cuOpt optimization problems","Troubleshoot cuOpt LP\u002FMILP problems including errors, wrong results, infeasible solutions, performance issues, and status codes. Use when the user says something isn't working, gets unexpected results, or needs help diagnosing issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1957,1960,1963,1964],{"name":1958,"slug":1959,"type":15},"Debugging","debugging",{"name":1961,"slug":1962,"type":15},"Mathematics","mathematics",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-23T05:43:32.51428",{"slug":1967,"name":1967,"fn":1968,"description":1969,"org":1970,"tags":1971,"stars":23,"repoUrl":24,"updatedAt":1975},"cuopt-model-mapper","map optimization problems to cuOpt models","Map interpreted optimization problems into cuOpt-native models for the fast path with minimal clarifying questions.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1972,1973,1974],{"name":20,"slug":21,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-30T05:29:19.401656",{"slug":104,"name":104,"fn":1977,"description":1978,"org":1979,"tags":1980,"stars":23,"repoUrl":24,"updatedAt":1989},"run cuOpt optimization in NemoClaw sandbox","Run cuOpt in the NemoClaw sandbox — probe\u002Fsmoke gates, prefer cancelable Python gRPC jobs, use legacy remote execution only when that API is unavailable, then vendored cuOpt skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1981,1984,1985,1986],{"name":1982,"slug":1983,"type":15},"CLI","cli",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":1987,"slug":1988,"type":15},"Simulation","simulation","2026-07-30T05:29:17.131364",{"slug":1991,"name":1991,"fn":1992,"description":1993,"org":1994,"tags":1995,"stars":23,"repoUrl":24,"updatedAt":2002},"generic-max-supply","plan supply chain models with cuOpt","Multi-period supply chain planning model: data files, BOM structure, variable\u002Fconstraint reference for the max-supply base model.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1996,1997,1998,1999],{"name":20,"slug":21,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"name":2000,"slug":2001,"type":15},"Supply Chain","supply-chain","2026-07-14T05:32:29.39577",{"slug":1642,"name":1642,"fn":2004,"description":2005,"org":2006,"tags":2007,"stars":23,"repoUrl":24,"updatedAt":2016},"orchestrate optimization data and solving","Coordinate uploaded data plus a natural-language question into interpretation, clarification, cuOpt solve, and a user-facing answer.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2008,2011,2014,2015],{"name":2009,"slug":2010,"type":15},"Automation","automation",{"name":2012,"slug":2013,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-14T05:32:59.680482",{"slug":1919,"name":1919,"fn":2018,"description":2019,"org":2020,"tags":2021,"stars":23,"repoUrl":24,"updatedAt":2028},"classify optimization and analytics requests","Classify whether a data-backed request is LP, MILP, QP, routing, or non-optimization analytics.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2022,2025,2026,2027],{"name":2023,"slug":2024,"type":15},"Analytics","analytics",{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-14T05:32:53.475243",{"slug":1899,"name":1899,"fn":2030,"description":2031,"org":2032,"tags":2033,"stars":23,"repoUrl":24,"updatedAt":2037},"route optimization requests to cuOpt solvers","Choose fast direct-to-cuOpt solve versus replayable or auditable model artifact mode.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2034,2035,2036],{"name":2009,"slug":2010,"type":15},{"name":9,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-30T05:29:18.162351",8,{"items":2040,"total":2195},[2041,2059,2077,2088,2100,2112,2125,2139,2152,2163,2177,2186],{"slug":2042,"name":2042,"fn":2043,"description":2044,"org":2045,"tags":2046,"stars":2056,"repoUrl":2057,"updatedAt":2058},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2047,2050,2053],{"name":2048,"slug":2049,"type":15},"Documentation","documentation",{"name":2051,"slug":2052,"type":15},"MCP","mcp",{"name":2054,"slug":2055,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":2060,"name":2060,"fn":2061,"description":2062,"org":2063,"tags":2064,"stars":2074,"repoUrl":2075,"updatedAt":2076},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2065,2068,2071],{"name":2066,"slug":2067,"type":15},"Containers","containers",{"name":2069,"slug":2070,"type":15},"Deployment","deployment",{"name":2072,"slug":2073,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":2078,"name":2078,"fn":2079,"description":2080,"org":2081,"tags":2082,"stars":2074,"repoUrl":2075,"updatedAt":2087},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2083,2086],{"name":2084,"slug":2085,"type":15},"CI\u002FCD","ci-cd",{"name":2069,"slug":2070,"type":15},"2026-07-14T05:25:59.97109",{"slug":2089,"name":2089,"fn":2090,"description":2091,"org":2092,"tags":2093,"stars":2074,"repoUrl":2075,"updatedAt":2099},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2094,2095,2096],{"name":2084,"slug":2085,"type":15},{"name":2069,"slug":2070,"type":15},{"name":2097,"slug":2098,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":2101,"name":2101,"fn":2102,"description":2103,"org":2104,"tags":2105,"stars":2074,"repoUrl":2075,"updatedAt":2111},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2106,2107,2108],{"name":1958,"slug":1959,"type":15},{"name":2097,"slug":2098,"type":15},{"name":2109,"slug":2110,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":2113,"name":2113,"fn":2114,"description":2115,"org":2116,"tags":2117,"stars":2074,"repoUrl":2075,"updatedAt":2124},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2118,2121],{"name":2119,"slug":2120,"type":15},"Best Practices","best-practices",{"name":2122,"slug":2123,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":2126,"name":2126,"fn":2127,"description":2128,"org":2129,"tags":2130,"stars":2074,"repoUrl":2075,"updatedAt":2138},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2131,2134,2137],{"name":2132,"slug":2133,"type":15},"Machine Learning","machine-learning",{"name":2135,"slug":2136,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":2140,"name":2140,"fn":2141,"description":2142,"org":2143,"tags":2144,"stars":2074,"repoUrl":2075,"updatedAt":2151},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2145,2148],{"name":2146,"slug":2147,"type":15},"QA","qa",{"name":2149,"slug":2150,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":2153,"name":2153,"fn":2154,"description":2155,"org":2156,"tags":2157,"stars":2074,"repoUrl":2075,"updatedAt":2162},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2158,2159],{"name":2069,"slug":2070,"type":15},{"name":2160,"slug":2161,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":2164,"name":2164,"fn":2165,"description":2166,"org":2167,"tags":2168,"stars":2074,"repoUrl":2075,"updatedAt":2176},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2169,2172,2173],{"name":2170,"slug":2171,"type":15},"Code Review","code-review",{"name":2097,"slug":2098,"type":15},{"name":2174,"slug":2175,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":2178,"name":2178,"fn":2179,"description":2180,"org":2181,"tags":2182,"stars":2074,"repoUrl":2075,"updatedAt":2185},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2183,2184],{"name":2146,"slug":2147,"type":15},{"name":2149,"slug":2150,"type":15},"2026-07-14T05:25:54.928983",{"slug":2187,"name":2187,"fn":2188,"description":2189,"org":2190,"tags":2191,"stars":2074,"repoUrl":2075,"updatedAt":2194},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[2192,2193],{"name":2009,"slug":2010,"type":15},{"name":2084,"slug":2085,"type":15},"2026-07-30T05:29:03.275638",496]