Overview
What this challenge is about.
Design the four-agent topology with explicit message contracts. Implement each agent as a separate LLM call with role-specific system prompts, tool access (web search for researcher, retrieval/recompute for fact-checker), and structured output schemas. Build a small eval harness: 30 questions with reference answers + acceptable-citation lists. Run the system, report accuracy + citation precision + USD cost per question + median time-to-answer. Compare against a single-agent baseline. Write a 4-page memo on when multi-agent is worth the cost and when it's not.
The Brief
What you'll do, and what you'll demonstrate.
Build a four-agent research assistant and quantify when the multi-agent topology beats a single-agent baseline on accuracy, citations, and cost.
Earning criteria — what you'll demonstrate
- Implement a multi-agent topology with role-specific prompts and tools
- Define structured message contracts between agents
- Evaluate multi-agent systems honestly against single-agent baselines
- Make a cost-benefit call on multi-agent architectures
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
AI Engineer
Building multi-agent topologies with LangGraph and shipping them with eval harnesses is the AI-engineer skill set that AI-agent and consulting orgs hire for in 2026.
This challenge sharpens
- multi-agent-orchestration
- langgraph
- llm-tool-use
Applied AI Scientist
Designing fair multi-vs-single-agent evaluations and writing the cost-benefit memo is the applied-AI work that product orgs need before scaling agents.
This challenge sharpens
- evaluation
- multi-agent-orchestration
- prompt-engineering
Prompt Engineer
Crafting role-specific prompts and structured contracts that survive multi-step coordination is core prompt-engineering work at agent-platform companies.
This challenge sharpens
- prompt-engineering
- llm-tool-use
- multi-agent-orchestration