Skip to main content
Back to challenges
Challenge

Build a Sanctions and Entity Resolution Agent with OpenAI Agents SDK and Qwen 3

Financial compliance teams struggle with fast entity resolution across sanctions lists and ambiguous ownership structures. Build an automated sanctions screening pipeline using OpenAI Agents SDK and Qwen 3 that cross-references complex ownership graphs and generates actionable risk scores.

CybersecurityHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Construct an AI agent that extracts entity relationship networks and resolves sanction evasion pathways across multi-jurisdiction watchlist feeds.

Delivery guide

How work is evaluated

Evaluation

Evaluates entity resolution accuracy and sanction risk scoring on synthetic ownership data.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

test_high_risk_flagging

Verify high risk score assigned to direct sanctioned matches

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

f1_score

F1 score for match detection • target: 0.95 • range: 0.85-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Implement fuzzy entity matching and graph-based ultimate beneficial ownership (UBO) extraction.

  • Utilize OpenAI Agents SDK with Qwen 3 to orchestrate entity resolution tools.

  • Deliver automated sanction risk scoring with traceable decision paths.

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic multi-jurisdictional entity profiles with obfuscated ownership networks.

Challenge setup

Evaluates entity resolution accuracy and sanction risk scoring on synthetic ownership data.

Challenge input

JSON containing entity profile and target watchlist record

OpenAI Agents SDK

Agent orchestration framework requirement

Qwen 3

Model assignment requirement

Evaluated output

JSON containing match_confidence, ubo_chain, and risk_score

Evaluation checks
  • Verify high risk score assigned to direct sanctioned matches
  • F1 score for match detection • target: 0.95 • range: 0.85-1
Configured evaluation evidence
  • Benchmark: Entity Resolution Benchmark v1
  • F1 Score target: 0.95
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • OpenAI · Required
  • Alibaba Cloud (Qwen) · Optional
Policy Serving
  • Qwen 3 · Optional

Evaluation contract

  • test_high_risk_flagging · Weight 1
  • f1_score · Weight 1

Metrics

  • f1_score · Target 0.95

Benchmarks

  • Entity Resolution Benchmark v1
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing entity profile and target watchlist record

    Output: JSON containing match_confidence, ubo_chain, and risk_score

    Execution harness: Python

Evaluation tools

  • OpenAI Agents SDK · agent_framework
  • Qwen 3 · llm

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Sanctions and Entity Resolution Agent with OpenAI Agents SDK and Qwen 3