aiengineering.guideaiengineering.guide

Zero to AI engineer · Stop 13 of 14 lessons

The three portfolio projects that get you hired

What this lesson covers

Everyone’s building a chatbot. Nobody’s reading them. This lesson names three project shapes that stand out because they demonstrate the skills teams actually need in 2026.

The outline

  1. Project one — the eval harness. Pick a public benchmark, build an eval harness for it, blog the results. Why this signals “can actually ship” faster than any demo.
  2. Project two — the RAG system with honest numbers. A retrieval system over a real corpus, with retrieval metrics and answer metrics reported separately.
  3. Project three — the agent that does one thing well. Not general-purpose. One tool, one workflow, ruthlessly scoped, with telemetry that shows it working.
  4. Scoping so you finish. The three-cut rule — cut everything that doesn’t fit in a weekend, then cut half of what remains.
  5. The README that gets read. Structure, screenshots, one paragraph on tradeoffs. What to put at the top.
  6. What not to build. The projects everyone submits, and why they all blur together.

Coming soon

In outline.

Outline

Enter to go · Esc to close