04. Week 16 — Daily Recall

Use this with 02_explainer.md and 03_study_material.md. Answer aloud before looking.

Monday

1. In the ship analogy, what are the course, the compass, the crew, the weather check, and the ship's log?
2. Retell the Sprint 1 to Sprint 5 failure story from memory.
3. What specifically separates a strong senior engineer from a Lead in this module?

Tuesday

1. Explain build vs buy vs fine-tune in under two minutes.
2. Give one example of a two-way door and one example of a one-way door in AI engineering.
3. What belongs in an ADR, and what should trigger a revisit?

Wednesday

1. What should be unit tested in an AI system?
2. What should be evaluated instead of unit tested?
3. Name the minimum production signals you would log for an LLM feature.

Thursday

1. What types of technical debt appear in AI systems beyond code debt?
2. What questions should AI code review ask that normal code review might miss?
3. How do you plan research-heavy sprint work without pretending uncertainty is gone?

Friday

1. When should a workflow stay manual for now?
2. Translate one technical decision separately for product, legal, and finance.
3. Why are principles context-dependent rather than universal rules?

Saturday

1. Without notes, recreate the failure-fix table from Chapter 6.
2. Pick one feature and explain its manual-to-automation roadmap.
3. State the four habits Module 17 assumes you already have.

Sunday

1. What is your personal default for API-first, and when would you override it?
2. Which principle from this week would most improve your current project?
3. Bridge question: how will MLOps make your preferred principles automatic?