Actuarial Climate Risk Scoring Agent with Pydantic AI and Type Safety
Actuaries navigating evolving climate risks and insurance claims require rigorous, type-safe data pipelines to compute risk exposure scores accurately. Build a Pydantic AI agent that ingests climate simulation vectors, validates exposure metrics against strict schema constraints, and outputs structured actuarial risk models for Singapore underwriters.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a production-grade, type-safe actuarial risk evaluation agent using Pydantic AI to assess climate risk exposure for Asian re-insurance portfolios.
How work is evaluated
Evaluates Pydantic AI type safety, validator enforcement, and numerical precision when calculating climate risk premiums.
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.
loss_math_validation
Verifies that Annualized Expected Loss = Asset Value * Flood Prob * Vulnerability Index
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
schema_compliance
Validation compliance score across model outputs • target: 1 • 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
Implement type-safe Python AI agents using Pydantic AI and Agent dependency injection
Define strict Pydantic schemas for climate modeling and catastrophic claim exposure
Integrate validation functions to enforce actuarial mathematical bounds automatically
Generate certified risk models aligned with Asian Actuarial Conference guidelines
Reference links and supporting material
Southeast Asian climate simulation data and commercial property risk exposure tables.
Challenge setup
Evaluates Pydantic AI type safety, validator enforcement, and numerical precision when calculating climate risk premiums.
Challenge input
JSON containing asset coordinates, flood probability, and asset value SGD
Pydantic AI
Ensures runtime type safety and strict schema validation for financial calculations
Evaluated output
JSON matching Pydantic model with asset_id, annualized_expected_loss, and risk_tier
- Verifies that Annualized Expected Loss = Asset Value * Flood Prob * Vulnerability Index
- Validation compliance score across model outputs • target: 1 • range: 0-1
- Benchmark: Type-Safe Agent Verification Benchmark
- Schema Compliance target: 1
- 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
- Pydantic AI · Required
- Aim · Optional
- Coplay AI · Optional
- Pydantic AI · Required
Evaluation contract
- loss_math_validation · Weight 1
- schema_compliance · Weight 1
Metrics
- schema_compliance · Target 1
Benchmarks
- Type-Safe Agent Verification Benchmark
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing asset coordinates, flood probability, and asset value SGD
Output: JSON matching Pydantic model with asset_id, annualized_expected_loss, and risk_tier
Execution harness: Python
Evaluation tools
- Pydantic AI · Type-Safe Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.