# Portfolio case study and resume worksheet
Original worksheet by Codex | September 28, 2026

## Case study structure
1. User and problem: one clear paragraph.
2. Baseline: what the non-AI or initial solution did.
3. Architecture: inputs, data, model, evaluation and delivery.
4. Evidence: actual measurements with the dataset and sample size.
5. Failure: an example that did not work and how you investigated it.
6. Tradeoff: a decision about quality, speed, cost or complexity.
7. Ownership: your contribution and how AI assistance was reviewed.
8. Links: demo, repository, evaluation instructions.

## Resume bullet formula
Built [specific feature] for [user/task] using [relevant implementation]; evaluated on [real dataset and sample size], achieving [actual measured result] versus [baseline], with [latency/cost/reliability context].

Keep placeholders until measured. Never turn a synthetic experiment into a claim of real customer impact. If the result is not an improvement, describe the insight honestly.

## Evidence map for one job
| Requirement | My evidence | Where a reviewer can see it | Gap |
|---|---|---|---|
| Programming / APIs | | | |
| LLM application | | | |
| Evaluation | | | |
| Delivery / operations | | | |
| Communication | | | |

## Final review
- Does the first half of the resume explain the relevant work clearly?
- Is every number traceable to a real measurement?
- Can all project links be opened without requesting access?
- Is confidential data absent?
- Can I explain every technology I list?
- Did I preserve my real job titles and dates?