Skip to main content
Loading navigation
Back to challenges

Claude Agents SDK Soft Market Premium Optimization

Commercial lines underwriters struggle to maintain margin discipline during soft market cycles with falling premium rates. Build a Claude Agents SDK workflow hosted on Baseten to forecast commercial property loss ratios and compute optimal rate adjustments.

Machine LearningHosted by Vera
Sign in free to start

What you are building

Deploy a Claude agent pipeline on Baseten that evaluates market rate movements and recommends calibrated renewal pricing.

How work is evaluated

Evaluation

Evaluates optimal policy rate pricing recommendations under simulated soft market competitive pressures.

Datasets

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

How submissions are scored

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

margin_discipline_check

Ensures expected combined ratio is <= target combined ratio

Binary check

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

Dimension 2

rate_optimality_score

Optimization score based on retention risk vs margin target • 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

  • Utilize the Claude Agents SDK to construct an actuarial pricing recommendation loop

  • Package and deploy model server functions onto Baseten infrastructure

  • Evaluate loss ratios and soft market pricing pressure to ensure statutory profitability

Resources

Dataset notes

CIAB Commercial Market Survey quarterly rate index values mapped to historic policy renewals.

Challenge setup

Evaluates optimal policy rate pricing recommendations under simulated soft market competitive pressures.

Challenge input

JSON policy history and market rate change metrics

Claude Agents SDK

Provides extended thinking capabilities for quantitative pricing reasoning

Baseten

Hosts and scales the underwriting evaluation models

Evaluated output

JSON containing recommended_rate_change_pct and target_combined_ratio

Evaluation checks
  • Ensures expected combined ratio is <= target combined ratio
  • Optimization score based on retention risk vs margin target • target: 0.95 • range: 0-1
Configured evaluation evidence
  • Benchmark: Actuarial-Pricing-Bench
  • Rate Optimality Score target: 95%
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Baseten · Required
  • AnythingLLM · Optional
  • Writer · Optional

Evaluation contract

  • margin_discipline_check · Weight 1
  • rate_optimality_score · Weight 1

Metrics

  • rate_optimality_score · Target 0.95

Benchmarks

  • Actuarial-Pricing-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON policy history and market rate change metrics

    Output: JSON containing recommended_rate_change_pct and target_combined_ratio

    Execution harness: Python

Evaluation tools

  • Claude Agents SDK · Agent Framework
  • Baseten · Inference Host Platform

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Claude Agents SDK Soft Market Premium Optimization