Ship an MVP RAG Knowledge Assistant for a Climate-Tech Startup
Overview
What this challenge is about.
As a 4-person team across a 6-week sprint, ship: (1) an ingestion pipeline for around 4,000 mixed PDFs and markdown files; (2) a vector store with documented chunking strategy; (3) a retrieval-augmented-generation (RAG) backend hitting a hosted large language model (LLM); (4) a small web app with sign-on integration (NextAuth or Clerk demo tier); (5) per-query logging and a tiny eval harness on 30 reference questions. Use standard software-engineering practice: Git branching, code review, CI tests, deployable infrastructure-as-code. Produce a 6-page engineering writeup and a 30-minute demo for the climate-tech founders.
The Brief
What you'll do, and what you'll demonstrate.
Ship a team-built MVP RAG knowledge assistant over a 4,000-document operations library with auth, logging, and an eval harness.
Earning criteria — what you'll demonstrate
- Ship a small AI product as a team using real software-engineering practice
- Design an ingestion + retrieval pipeline for mixed-format documents
- Operate a basic eval loop alongside the product
- Communicate engineering decisions to a non-engineering audience
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
Shipping a deployed RAG product as a team is the literal day-one job description for an AI engineer at any LLM-product startup.
This challenge sharpens
- retrieval-augmented-generation
- software-engineering-for-ai
- python
Machine Learning Engineer
Pipeline design, CI discipline, and eval-in-CI mirror the MLE's daily craft on production ML systems.
This challenge sharpens
- software-engineering-for-ai
- ci-cd
- python
AI Solutions Architect
Owning the ingestion-plus-retrieval-plus-eval design across a team is the architect's contribution at scale.
This challenge sharpens
- vector-databases
- retrieval-augmented-generation
- team-collaboration