Orchestrate Multi-Agent Actuarial Risk Rating using AutoGen & Hydra
Actuaries must evaluate complex commercial property catastrophe risks using diverse datasets. Orchestrate an AutoGen multi-agent system with Hydra configuration management to dynamically simulate risk models, parse GAIP case scenarios, and output audited premium recommendations with less than 2% calculation variance across multi-run trials.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an AutoGen multi-agent system managed by Hydra configurations to evaluate commercial actuarial risk and catastrophe exposure.
How work is evaluated
Evaluates the consistency, accuracy, and auditability of premium recommendations across stochastic multi-agent deliberation runs.
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.
consensus_achieved
Verifies that AutoGen agents reached consensus within configured max turns.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
premium_variance
Variance of technical premium calculations across 5 runs. • target: 0.015 • range: 0-0.05
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Master AutoGen multi-agent conversation patterns for actuarial consensus
Use Hydra for reproducible execution configurations and experiment tracking
Implement actuarial risk calculation tools within AutoGen agent routines
Evaluate multi-agent deliberation workflows for stability and variance reduction
Reference links and supporting material
Synthetic catastrophe risk dataset featuring Singapore commercial maritime infrastructure portfolios.
Challenge setup
Evaluates the consistency, accuracy, and auditability of premium recommendations across stochastic multi-agent deliberation runs.
Challenge input
Hydra dynamic YAML config + property portfolio details
AutoGen
Enables specialized multi-agent conversation and consensus flows.
Hydra
Provides dynamic hierarchical configuration for multi-agent parameters.
Evaluated output
JSON actuarial opinion summary with calculated technical premium
- Verifies that AutoGen agents reached consensus within configured max turns.
- Variance of technical premium calculations across 5 runs. • target: 0.015 • range: 0-0.05
- Benchmark: ActuaryBench
- Premium Variance target: 0.015
- 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
- AutoGen · Required
- Hydra · Optional
- Ax (Adaptive Experimentation) · Optional
Evaluation contract
- consensus_achieved · Weight 1
- premium_variance · Weight 1
Metrics
- premium_variance · Target 0.015
Benchmarks
- ActuaryBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: Hydra dynamic YAML config + property portfolio details
Output: JSON actuarial opinion summary with calculated technical premium
Execution harness: Python
Evaluation tools
- AutoGen · Multi-Agent Orchestration
- Hydra · Configuration Management
Recipe freeze marker
No recipe freeze time is recorded.