Still Figuring It Out: Honest Notes from Building with Claude Code and Job Hunting

Everyone is talking about how AI will transform software development and hiring. But what does it actually look like when an experienced tech leader ditches traditional routines and relies on AI tools to build products and navigate a modern job search?

Where do classic agile processes break down when coding with AI? How much can you actually trust AI feedback on your CV and company research? And what decisions should you deliberately keep out of the algorithm’s hands?

For developers and engineering leaders, the shift toward agentic coding tools like Claude Code can feel disorienting—traditional backlogs, Kanban boards, and PR workflows often don't survive contact with rapid AI prototyping. At the same time, integrating AI into a job search can deliver brilliant insights one day and costly dead ends the next.

This talk cuts through the online highlight reels to offer an honest, practitioner's account of shipping two real applications with Claude Code, using AI across a live job search, and facing the unsolved questions about how software teams will work in the future.

We'll explore:

- Rapid prototyping with Claude Code: what worked, what broke, and shipping two apps in weeks

- Why textbook agile workflows, PR cycles, and backlogs quickly fell apart—and what replaced them

- The reality of AI-assisted job hunting: impartial CV matching, expensive research traps, and where human judgment must stay in control

- How individual developer velocity impacts team structure and future software delivery

- Practical takeaways and unresolved challenges from the front line of building with AI

You don't need theoretical hype to understand how AI is changing development. Come along to hear Graham share practical, candid lessons from the tools, the workflow experiments, and the realities of modern technical job hunting.