Adnoc Gas EPC Supply Chain Orchestration with LangChain & Claude 4.1 Opus
Managing procurement across $8.2bn in Adnoc Gas projects requires tracking multi-tier supplier delays and material bottlenecks. Build a LangGraph stateful agent pipeline powered by Claude 4.1 Opus to analyze contractor schedules, update procurement risk states, and dynamically trigger mitigation options. Reach an evaluation accuracy of 92% on supply bottleneck identification.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a stateful EPC supply chain agent using LangChain/LangGraph and Claude 4.1 Opus.
How work is evaluated
Evaluates EPC project delay bottleneck identification and state graph execution.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Correct Critical Path Selection
Verifies that identified package has maximum delay impact.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Bottleneck Identification Precision
Accuracy of critical path delay detections. • target: 0.92 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build complex graph workflows using LangChain and LangGraph
Integrate Anthropic Claude 4.1 Opus for high-reasoning supply chain analysis
Track multi-contractor dependencies across Adnoc project packages
Generate automated mitigation strategies for critical path delays
Reference links and supporting material
Synthesized procurement schedule and delay dataset based on $8.2bn Adnoc Gas awards.
Challenge setup
Evaluates EPC project delay bottleneck identification and state graph execution.
Challenge input
JSON object with packages array, contractor_status, lead_times
LangChain
Required framework for building stateful graph agents.
Claude 4.1 Opus
High-reasoning intelligence engine for complex schedule analysis.
Evaluated output
JSON object with bottleneck_package_id, delayed_days, mitigation_action
- Verifies that identified package has maximum delay impact.
- Accuracy of critical path delay detections. • target: 0.92 • range: 0-1
- Benchmark: ProcureBench
- Bottleneck Identification Precision target: 0.92
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Langchain · Optional
- Dagster · Optional
- Claude 4.1 Opus · Required
- Langchain · Optional
Evaluation contract
- Correct Critical Path Selection · Weight 1
- Bottleneck Identification Precision · Weight 1
Metrics
- Bottleneck Identification Precision · Target 0.92
Benchmarks
- ProcureBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with packages array, contractor_status, lead_times
Output: JSON object with bottleneck_package_id, delayed_days, mitigation_action
Execution harness: Python
Evaluation tools
- LangChain · Agent Architecture
- Claude 4.1 Opus · Foundation Model
Recipe freeze marker
No recipe freeze time is recorded.