Build an Intelligent Crude Export Clearance Agent with LlamaIndex and Arcee AI
Iraq's approval of a three-month mechanism for crude oil exports via local and international trading firms introduces temporary regulatory parameters and dynamic counterparty verification rules. Developers must build a domain-specific retrieval-augmented generation (RAG) system capable of parsing complex bilingual petroleum sales agreements, official gazettes, and SOMO (State Organization for Marketing of Oil) allocation notices. Using LlamaIndex and Arcee AI, you will index structured and unstructured export rules, implement contextual routing, and orchestrate an agent that verifies whether proposed local and foreign trading firms meet interim capitalization, compliance, and lift volume quotas.
What you are building
The core problem, expected build, and operating context for this challenge.
Implement a bilingual crude oil export clearance agent using LlamaIndex contextual retrieval and Arcee AI workflow automation.
How work is evaluated
Evaluates accuracy of trade firm qualification and quota clearance under the 3-month mechanism.
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.
check_regulatory_compliance
Ensure firm meets minimum capital and quota constraints.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
clearance_accuracy
Accuracy across 50 simulated firm evaluation scenarios. • target: 0.95 • 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
Master contextual RAG indexing for complex bilingual petroleum trade regulatory documents.
Implement multi-document routing and metadata filtering using LlamaIndex.
Integrate Arcee AI workflows to synthesize compliance verdicts and verification records.
Reference links and supporting material
Synthetic dataset of 50 GCC crude trading company profiles and interim 3-month export rules.
Challenge setup
Evaluates accuracy of trade firm qualification and quota clearance under the 3-month mechanism.
Challenge input
JSON containing firm_profile and requested_lift_volume_bpd
LlamaIndex
Provides robust data framework for complex RAG operations.
Arcee AI
High efficiency specialized domain models for trade operations.
Evaluated output
JSON with clearance_status, violated_clauses, and recommended_allocation
- Ensure firm meets minimum capital and quota constraints.
- Accuracy across 50 simulated firm evaluation scenarios. • target: 0.95 • range: 0-1
- Benchmark: GCC-Energy-Trade-Bench
- Clearance Accuracy target: 0.95
- 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
- Arcee AI · Required
- Llama Index · Optional
- Exa · Optional
Evaluation contract
- check_regulatory_compliance · Weight 1
- clearance_accuracy · Weight 1
Metrics
- clearance_accuracy · Target 0.95
Benchmarks
- GCC-Energy-Trade-Bench
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing firm_profile and requested_lift_volume_bpd
Output: JSON with clearance_status, violated_clauses, and recommended_allocation
Execution harness: Python
Evaluation tools
- LlamaIndex · Orchestration & RAG Indexing
- Arcee AI · SLM Inference
Recipe freeze marker
No recipe freeze time is recorded.