Build a Small-Bank Liquidity and Counterparty Risk Monitor with LangGraph
Regional banking stress requires continuous telemetry analysis of liquidity metrics and deposit outflows. Build a LangGraph stateful agent pipeline to model bank systemic risk and alert treasury teams to potential default cascades.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct a stateful multi-step agent pipeline using LangGraph to monitor liquidity metrics and model counterparty risk.
How work is evaluated
Evaluates bank failure contagion predictions under stress testing scenarios.
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.
Critical Liquidity Threshold Flag Test
LCR under 0.85 with high uninsured deposits must trigger CRITICAL state.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Contagion Early Warning Precision
Accuracy of bank failure warnings. • target: 0.95 • range: 0.9-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Construct a stateful execution graph in LangGraph with conditional routing based on liquidity threshold breach.
Analyze bank regulatory call reports, core deposit stability, and uninsured deposit ratios.
Simulate counterparty contagion risks across interbank lending connections.
Trigger automated emergency treasury warnings when failure probabilities rise above threshold.
Reference links and supporting material
Synthetic financial report dataset detailing small bank liquidity and deposit flight telemetry.
Challenge setup
Evaluates bank failure contagion predictions under stress testing scenarios.
Challenge input
JSON containing core_deposits, uninsured_deposits_pct, liquidity_coverage_ratio, uncollected_loans
LangChain
Provides core LLM wrappers and tool structures.
LangGraph
Manages persistent stateful execution nodes and conditional transitions.
Evaluated output
JSON with failure_risk_stage ('HEALTHY', 'WATCH', 'CRITICAL'), contagion_index
- LCR under 0.85 with high uninsured deposits must trigger CRITICAL state.
- Accuracy of bank failure warnings. • target: 0.95 • range: 0.9-1
- Benchmark: BankLiquidityBench
- Contagion Early Warning Precision 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
- LangGraph · Required
- Langchain · Optional
- LangChain · Optional
- LangGraph · Required
- Langchain · Optional
- LangChain · Optional
Evaluation contract
- Critical Liquidity Threshold Flag Test · Weight 1
- Contagion Early Warning Precision · Weight 1
Metrics
- Contagion Early Warning Precision · Target 0.95
Benchmarks
- BankLiquidityBench
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing core_deposits, uninsured_deposits_pct, liquidity_coverage_ratio, uncollected_loans
Output: JSON with failure_risk_stage ('HEALTHY', 'WATCH', 'CRITICAL'), contagion_index
Execution harness: Python
Evaluation tools
- LangChain · Agent Framework
- LangGraph · State Graph Engine
Recipe freeze marker
No recipe freeze time is recorded.