FAQ

FAQ

Q: Why is there no AI in the first month? An agent's memory, tool results, and step records live in a database. You can't build a reliable agent without understanding the database well. Also, most interviews are system design + database questions.

Q: Isn't two languages too much? Should we stick to one? The target profile (AI Platform Engineer) requires both worlds: Python is the home of the LLM ecosystem, the JVM is the home of orchestration and security. Splitting later is a painful refactor. See ADR-0001.

Q: If we both learn the same thing, isn't that a waste of time? No — the real waste is being two half engineers six months from now. Learning is not divided, implementation is. You close the part you didn't implement through review and the teaching session.

Q: What if we fall a week behind? Don't shorten the plan, shift it. Every month builds on the previous one; a half-finished month rots the next one too.

Q: Will we never learn deep learning? You'll learn it at a conceptual level: what it is, why it works, when it's used. Going into the math hurts this profile — the engineer who trains models and the engineer who runs models are different professions.

Q: Can we have AI write code for us? The rule is simple: can you defend every line that goes into the repo? If yes, no problem. If no, that line shouldn't go in. See Learning Method.

Q: Why is the repo private at first? It's in the skeleton stage until the Month 6 launch. Going public early means spending a first impression on a half-finished project. It's flipped to public at launch.

Q: Is this project really "one of a kind in the market"? No, and it doesn't need to be. The realistic goal: be one of the best open-source tools in a narrow gap like agent evaluation and observability. That's what gets you hired.

Q: What will we be at the end of six months? A strong Junior / Junior-to-Mid AI Platform Engineer with a clear specialization direction and a proven project. Not "Senior" — but clearly ahead of a directionless candidate.