# FDE System Decomposition Cheat Sheet

One-page prep card for Palantir-style decomp rounds and AI-lab applied cases.

## C.A.S.E. spine (say this out loud)

1. **Clarify** — success metric, users, data sources, security class, timeline, non-goals
2. **Architect** — thin end-to-end path from source system → operator workflow
3. **Solve the Delta** — what the product does not do OOTB; glue you build first
4. **Evaluate** — proof plan (evals / UAT / adoption) + Day-2 owner

## First five minutes questions

- What does "done" mean in a number or operator behavior?
- Where is the system of record today (and how dirty is it)?
- Who is the champion — and who can veto?
- What security / residency constraints are hard vs soft?
- What happens if we ship nothing for 30 days?

## Thin vertical slice checklist

- [ ] One persona
- [ ] One workflow
- [ ] Real (even incomplete) data path
- [ ] Measurable outcome
- [ ] Explicit non-goals
- [ ] Instrumentation / eval gate
- [ ] Rollback or human escalation path

## Instant fail modes

- Jump to solutions before constraints
- Boil-the-ocean architecture with no 1–2 week slice
- Ignore permissions, data quality, or political veto players
- End at a diagram with no evaluation plan
- Only speak engineer-to-engineer jargon

## AI-specific extras

- Separate retrieval failures from generation failures
- Keep irreversible writes deterministic or human-gated
- Set latency + cost ceilings before expanding autonomy
- Require citations / abstain when context is weak

## Companion links

- https://fdehub.dev/interviews/fde-case-study-decomp
- https://fdehub.dev/interviews/palantir-fde-interview-guide
- https://fdehub.dev/learn/business-to-ai-translation-matrix
- https://fdehub.dev/tools/readiness