Developer Experience in the age of AI

for more information visit https://salaboy.com

Transcript

3. Inner Loop → Outer Loop Developer-centric → Team collaboration
4. 4 The Inner Loop (before coding agents) code → build → test → debug This tasks allow developers to learn the in-and-outs of their applications
5. 5 Outer Loop Review PR → Merge to codebase → Release Collaboration happens when resolving merge conflicts based on team decisions and ADRs
6. 6 Coding Agents (automating the inner loop)
8. Coding Agent Tools Memory Workflows Large Language Model Automates the Inner Loop Tasks in a generic way Guardrails
9. 9 So what changed?
10. 10 prompt → coding agent → code Automated inner loop, no learnings It automates the coder but not the engineer Most of my time is now spent reviewing what was generated Requirements need to be expressed as prompts
11. Multiple parallel inner loops But it is true, I can do multiple things in parallel
12. 12 The final leap of faith!
13. 13 How do we make coding agents more efficient?
14. 14 Prompt Engineering MCP / Skills Conventions So it actually do what you want it to do So it does integrate with your organization systems
15. 15 Let’s be practical Let’s look at some tools
16. Demo #1 Building blocks to create your own Coding Agents
17. Demo #2 The YOLO mode
18. Demo #3 How about pair programming and collaborations?
19. Demo #4 If you don’t measure, how do you know it is better?
20. Demo #5 Tracking Intent to Code
21. The developers winning with coding agents … are the ones who redesigned their inner loop — with the same rigor they'd apply to any system they build. Which takes a lot of seniority to do
22. Companies winning with coding agents … are the ones that managed to standardize the new inner loops and solved the collaborations involved in the outer loops I haven’t seen many of this 🦄
23. Thanks @salaboy / salaboy.com Diagrid.io