Full Walkthrough: Workflow for AI Coding
summary
A concrete workflow shows ambiguous briefs turned into a PRD, then into local markdown issues split as traceable bullets so agents can grab independent work; an automated loop runs Claude Code (with Docker sandbox), uses the grill me skill to align requirements, runs an AFK implementer that writes failing tests and uses TDD to pass them, commits code and runs type checks — the demo spends ~25k tokens, produces commits and 284 tests, and yields runnable changes that require human QA.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:14The Thesis of AI Engineering
- 4:20Phase 1: Research & Prototyping
- 12:45Phase 2: The Grill Session
- 22:10Phase 3: Writing the PRD
- 35:50Phase 4: Slicing Work into Issues
- 48:15Phase 5: Implementation with AI Agents
- 1:05:30Phase 6: Human-in-the-Loop Review
- 1:18:45Phase 7: Deployment & Monitoring
- 1:28:10Designing Codebases for AI Effectiveness
- 1:34:06Final Takeaways & Summary
related talks
01h 06mHow I Tamed Claude
26mI’m Not Writing Code Anymore—Now What? | The 7-Step Workflow
01h 40mProcess files: Blueprints for Agentic AI
31mThe Copilot Wrote It but Who Saves It: Surviving the AI Code Tsunami
01h 01m10 tips to level up your ai-assisted coding
30mBuilding Tools with Tools: Modern AI-Driven Development
28mHow To Use Spec-Driven Development for Production Workflows
01h 00mHow I Tamed Claude