0 of 2 done
From the AI & ML course. Split by the three career tracks, but weighted toward AI engineering — the deepest-built part of the folder and the applied area with the most shipped work behind it.
Files
- AI Engineering — LLMs, prompting, RAG, tools/MCP, agents, evals, safety
- ML & Stats Fundamentals — bias/variance, evaluation metrics, leakage, A/B testing
What these interviews actually test
For AI engineering roles the differentiator is almost never model knowledge — it’s whether you treat an LLM system as software with a probabilistic component or as a demo. The questions that separate people are about evaluation, failure modes, and cost, not about prompt tricks.
If you can say “here’s my eval set, here’s the success rate, here’s what changed when I swapped models,” you’re ahead of most candidates.
Related
- AI & ML course · AI engineer track
- Project Ideas — the RAG-over-this-vault and MCP-server projects are directly interview material
- Databases — vector search sits in the same conversation
Jump to a subject
Software engineering
Infrastructure & security
Computer science
Hardware & engineering