Skip to main content
Back to challenges
Challenge

Pydantic AI Reinsurance M&A Portfolio Integration Engine

Merging reinsurance brokers must reconcile conflicting treaty structures and binder limits across disparate databases with total type safety. Implement a Pydantic AI agent system with strict schema validation to harmonize acquired portfolio schedules and flag overlapping treaty coverage.

Machine LearningHosted by Vera
Sign in free to start
Challenge brief

What you are building

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

Utilize Pydantic AI structured outputs and dependency injection to standardize acquired reinsurance binder records.

Delivery guide

How work is evaluated

Evaluation

Evaluates type validation enforcement and identification of overlapping treaty exposure layers.

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

pydantic_validation

Ensures model output validates without Pydantic ValidationError

Binary check

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

Dimension 2

field_extraction_f1

F1 score of extracted treaty field attributes • target: 0.95 • range: 0-1

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

  • Build a Pydantic AI agent featuring strong type validation for treaty mapping

  • Utilize Pydantic AI dependency injection for database integration during processing

  • Detect and log duplicate or conflicting excess-of-loss binder layers across merged portfolios

Resources and assets

Reference links and supporting material

Dataset notes

Dataset of 400 unformatted reinsurance treaty schedules from BMS Re / Willis Re merger files.

Challenge setup

Evaluates type validation enforcement and identification of overlapping treaty exposure layers.

Challenge input

Unstructured string of binder terms

Pydantic AI

Ensures Python data validation and structured outputs for agent pipelines

Evaluated output

Validated JSON matching TreatySchema with overlap flags

Evaluation checks
  • Ensures model output validates without Pydantic ValidationError
  • F1 score of extracted treaty field attributes • target: 0.95 • range: 0-1
Configured evaluation evidence
  • Benchmark: Schema-Extract-Bench
  • Field Extraction F1 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
  • Pydantic AI · Required
  • LocalAI · Optional
Policy Serving
  • DeepSeek V3 · Optional
Orchestration
  • Pydantic AI · Required

Evaluation contract

  • pydantic_validation · Weight 1
  • field_extraction_f1 · Weight 1

Metrics

  • field_extraction_f1 · Target 0.95

Benchmarks

  • Schema-Extract-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: Unstructured string of binder terms

    Output: Validated JSON matching TreatySchema with overlap flags

    Execution harness: Python

Evaluation tools

  • Pydantic AI · Type-Safe Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Pydantic AI Reinsurance M&A Portfolio Integration Engine