Lightening Talks on AI-Native Development

How AI Fixed a Terrible Legacy Codebase in 5 Days

AI agents are great at small, modular projects, but they struggle when faced with sprawling, messy code. The real limitation isn’t just the model size; it’s the lack of discipline to structure code and manage complexity.

As a technical lead, I’ve spent years coaching teams to form habits that improve code quality without slowing delivery. When I noticed AI agents run into the same issues as undisciplined teams, I asked: what if we built workflows that enforced those habits for AI?

The result was a breakthrough: we finally refactored a legacy codebase we’d avoided for years. In just five days, AI cut our operating costs by 70% a job that would have taken months of engineering effort.

In this session, I’ll share the setup, the practices behind it, and how you can adapt them to make AI not just flashy, but genuinely useful in your own complex, real-world codebases.

The Impact of GenAI on Development and the SDLC

Software development is entering a new wave of automation—one that mirrors the shift we saw with DevOps. Agentic AI tools are rapidly reshaping workflows, introducing new practices, and changing how teams think about the software delivery lifecycle.

In this talk, we’ll explore the emerging patterns behind these tools, what they mean for day-to-day development, and how to navigate the evolving landscape of AI Native Development with confidence.

AI Native Dev Meetup at InfoQ