# Technical ML project presentation outline

Build 10 to 12 slides for 30 minutes of content and reserve 15 minutes for questions.

1. **Thesis**: problem, stakes, your ownership, measured result
2. **System context**: users, constraints, prior approach, failure cost
3. **Decision 1**: alternatives, evidence available then, choice
4. **Architecture**: only the components needed to understand the decisions
5. **Technical deep dive**: one mechanism, bottleneck, or derivation
6. **Decision 2**: experiment or implementation trade-off
7. **Failure**: wrong assumption, signal, response
8. **Evaluation**: offline, online, slices, uncertainty, guardrails
9. **Launch and operations**: monitoring, fallback, ownership
10. **Impact**: observed outcome, attribution boundary, second-order effects
11. **What changed in your judgment**
12. **Decision summary**: three claims the audience should remember

## Delete these slides

- A resume timeline
- A literature survey unrelated to your decisions
- A dense architecture with every service
- A metric dashboard without a decision
- A team-org chart used to avoid stating personal ownership
- A victory lap with no failed attempt