Point72 Machine Learning Engineer Interview Experience — Reviewing a RAG Pull Request Instead of LeetCode

The role was MLE on the GenAI team. This was the round after the recruiter call. On the recruiter call they had said it would be a LeetCode-style interview, and it turned out to be completely different...

The first half went through the projects on my resume and the details of each project. My answers weren't very smooth. From now on I need to have my resume details prepared for every interview.

The second half was coding. They gave me a piece of code from a RAG system and told me to pretend it was a PR and find the problems in it. After we discussed the problems, I had to write test cases based on them, and the syntax didn't have to be exactly right. But I'm really not very familiar with writing tests, so I think there will probably be no follow-up.

async def answer(question, user):
    docs = await retrieve(question, top_k=8)
    visible = [doc for doc in docs if can_read(user, doc)]
    try:
        result = await internal_model(question, visible)
        if not result.citations:
            return {"status": "abstained"}
    except TimeoutError:
        result = await external_model(question, docs)
        if not result.citations:
            return {"status": "abstained"}
    await audit.write({
        "user_id": user.id,
        "question": question,
        "answer": result.text,
    })
    return {
        "status": "answered",
        "text": result.text,
        "citations": result.citations,
    }