Drawing from 2 years of hands-on experience with AI coding agents since building Kaguya (an open-source ChatGPT plugin), I'll demonstrate why software engineering fundamentally hasn't changed - it's always been about context engineering.
I'll share 7 practical techniques that have helped me write thousands of lines of production code with AI agents:
- Working with Git and GitHub CLI through agents
- Saving context while maximizing productivity
- Navigating unfamiliar codebases
- Completing the write-test cycle
- Doing research on unfamiliar topics
- Pair debugging with AI agents
- Using voice-to-text for faster development
Speaker Bio
- Name: YK Sugi
- Mini-bio: Developer Experience Manager at Sourcegraph. Also the creator of an open-source coding assistant with ~1200 stars on GitHub. Founder of CS Dojo, a programming education YouTube channel with 1.9M+ subscribers.
- Picture for slides:

Drawing from 2 years of hands-on experience with AI coding agents since building Kaguya (an open-source ChatGPT plugin), I'll demonstrate why software engineering fundamentally hasn't changed - it's always been about context engineering.
I'll share 7 practical techniques that have helped me write thousands of lines of production code with AI agents:
Speaker Bio