Build the application

## Build
Score and prioritize financial crime investigative queues using configurable fictional business rules. Map case attributes to investigator specialization and capacity to produce transparent operational rosters.

## User
AML Investigation Team Lead

## Workflow
1. Ingest batches of pending fictional financial crime case files containing entity risk, value, and alert types.
2. Calculate deterministic priority scores using weighted factors for corporate exposure, cross-border tags, and statutory SLA limits.
3. Evaluate investigator roster metrics including active load, jurisdiction certifications, and current backlog.
4. Generate a balanced investigation dispatch roster with explicit routing rationales and escalation markers.

## Interface
Investigative operations dispatch console showing backlog queues, investigator capacity gauges, and assignment logs.

## Data
Construct a synthetic JSON fixture of fictional alert records and investigator profiles.
Starter examples below are synthetic. They are not downloaded records.

## Source
[OpenSanctions data toolkit](https://raw.githubusercontent.com/opensanctions/yente/main/README.md)
Employ FollowTheMoney entity property types (jurisdiction, amount, currency, entity schema) for uniform alert attributes.

## Tools
langgraph, pydantic, fastapi.
LangGraph manages stateful queue routing loops and conditional dispatcher nodes efficiently.
Local or hosted LLM with tool-calling capabilities for routing decisions.

## Deliverables
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

## Acceptance
- Calculates deterministic priority rank adhering to SLA deadline and multi-factor risk weighting.
- Respects investigator caseload caps during automated assignment passes.
- Provides clear reason codes explaining why a case was routed to a specific analyst.
- Returns full priority queue JSON payload with valid ISO timestamps and routing keys.

## Starter Examples
```json
[
  {
    "input": {
      "cases": [
        {
          "case_id": "CASE-901",
          "entity_type": "Company",
          "jurisdiction": "ky",
          "aggregate_amount": 750000,
          "days_to_sla": 2
        },
        {
          "case_id": "CASE-902",
          "entity_type": "Person",
          "jurisdiction": "us",
          "aggregate_amount": 12000,
          "days_to_sla": 14
        }
      ],
      "investigators": [
        {
          "id": "inv_1",
          "name": "Elena Vance",
          "active_cases": 2,
          "max_capacity": 5,
          "specialties": [
            "offshore",
            "corporate"
          ]
        },
        {
          "id": "inv_2",
          "name": "Marcus Reid",
          "active_cases": 4,
          "max_capacity": 5,
          "specialties": [
            "retail",
            "structuring"
          ]
        }
      ]
    },
    "expected_output": {
      "dispatches": [
        {
          "case_id": "CASE-901",
          "assigned_to": "inv_1",
          "priority_tier": "critical",
          "score": 95,
          "routing_reason": "SLA expiring in 2 days and offshore corporate jurisdiction match."
        },
        {
          "case_id": "CASE-902",
          "assigned_to": "inv_2",
          "priority_tier": "low",
          "score": 30,
          "routing_reason": "Low value domestic person alert within safe SLA threshold."
        }
      ],
      "unassigned_cases": []
    }
  },
  {
    "input": {
      "cases": [
        {
          "case_id": "CASE-903",
          "entity_type": "Company",
          "jurisdiction": "pa",
          "aggregate_amount": 1500000,
          "days_to_sla": 1
        }
      ],
      "investigators": [
        {
          "id": "inv_3",
          "name": "Sarah Chen",
          "active_cases": 5,
          "max_capacity": 5,
          "specialties": [
            "offshore"
          ]
        }
      ]
    },
    "expected_output": {
      "dispatches": [],
      "unassigned_cases": [
        {
          "case_id": "CASE-903",
          "priority_tier": "critical",
          "score": 98,
          "status": "queued_capacity_exceeded",
          "routing_reason": "Investigator capacity full; requires supervisor rebalancing."
        }
      ]
    }
  }
]
```

## Safety
Utilizes purely synthetic scenarios with no real names, simulated routing only, and no automated regulatory reporting.
Use mock external writes. Require human review before any real deployment.