Design, prompt, tool-enable, and benchmark a production-ready AI research assistant.
Milestones (8)
- 01Define the Assistant Persona & Scope
Define domain boundaries, tone, role, and audience for the research assistant.
- 02Prompt Foundations & Constraints
Transform generic queries into structured, constraint-enforced system prompts.
- 03Structured Outputs (JSON Schema)
Enforce rigid JSON outputs for seamless downstream software integration.
- 04Information Retrieval & Synthesis
Design fact extraction pipelines that synthesize multiple technical sources.
- 05Tool Use & Function Calling Basics
Enable the assistant to query external APIs and calculation tools.
- 06Model Context Protocol (MCP) Simulation
Integrate MCP servers for structured context injection.
- 07Evaluation, Guardrails & Benchmarking
Implement automated safety filters, hallucination checks, and benchmark test suites.
- 08Final Capstone Showcase & Mentor ReviewMentor Review
Bundle all deliverables and submit for human mentor evaluation and certification.