Custom software

Where an engineer's work moves when AI drafts the code

1. Understand the problem
2. Specify
3. Generate
4. Review
5. Verify
6. Ship and measure

An intelligent app in layers: what fails in each, and what catches it

1. Application
2. Model
3. Retrieval (RAG)
4. Data

Every tip here is one step of the same loop, repeated on small projects

1. Pick one goal
2. Learn one concept
3. Build it yourself
4. Break it, then debug
5. Commit and test
6. Get it reviewed

What an AR app does on every frame, 60 times a second on a phone

1. Sense
2. Track
3. Understand
4. Anchor
5. Render

The software development life cycle, and what each phase must produce

1. Discovery
2. Requirements
3. Design
4. Build and test
5. Release
6. Operate and maintain

A first machine learning project in Python, from setup to a hosted model

1. Set up
2. Load and look
3. Split first
4. Train a pipeline
5. Test once
6. Go further

Where each kind of automated test runs, from commit to production

1. Commit
2. Pull request
3. Preview build
4. Staging
5. Production