aichallengesafrica
Take a challenge
HomeChallengesRepsProblemsLeaderboardTake a challenge
TRACK 04

AI + RAG

Chunk, embed, retrieve, cite — the parts of an AI feature that actually get graded in an interview.

6 challenges2 free2h of build time2 graded by running your code

RAG and Model APIs

Chunk it, rank it, ground it — then prove the answer came from the document.

  1. Chunk a DocumentSplit a document into retrievable pieces. No API key, no network — this is string handling, and it is where most RAG apps are already broken.Beginner · 15 min
  2. Rank by SimilarityCosine similarity, top-k, and a score floor — the retrieval half of RAG, written out by hand so you know what the vector database is doing for you.Beginner · 15 min
  3. Tool UseGive a model a real function to call, then handle the round trip — including the case where it asks for something that does not exist.Intermediate · 20 min · Pro
  4. Stream the AnswerPaint tokens as they arrive, and make a second question cancel the first. This is the whole difference between an AI feature that feels fast and one that feels broken.Intermediate · 20 min · Pro
  5. RAG Chat with CitationsThe whole pipeline end to end: upload a .txt, chunk and embed it, answer a question from it, and cite which chunks you used.Advanced · 25 min · Pro
  6. Refuse and EvaluateMake the thing say "I do not know" — then build the eval set that proves it still does after you change the prompt.Advanced · 25 min · Pro