Year-End Reflections: The Last 18 Months, and What Comes Next

As the year comes to a close, I’ve been reflecting on the last 18 months since leaving AWS. It’s been a long, focused chapter of deep work, and it’s now setting the stage for an exciting and very different year ahead.

Before I left, I had an incredible experience at AWS, working with exceptional people on hard problems and building networking infrastructure that supported AI workloads at massive scale. I’m deeply grateful for that chapter and for the people I learned from there.

That experience sharpened my curiosity about where AI was headed and what it might mean for how software gets built. In July of last year, I decided to leave and pursue a startup, believing AI was about to change software development in a real way and that the SDLC itself was overdue for a rethink.

Starting where the friction was

I started at the top of the stack with agentic AI. For more than a year, I spent nearly every day building software with AI, testing what it took to reach production-quality results, and learning quickly where things worked and where they broke.

Stepping back, this has been one of the most personally rewarding periods of my career. I had the space to go deep and learned a great deal about designing agentic systems, using AI to write production software, and what it really takes to deploy AI systems reliably.

Some of the ideas I was working through later became mainstream and showed up in products many people now use daily, including tools like Claude Code, Kiro, Kilo, and OpenCode.

The bigger problem was underneath

Along the way, something became clear. Automating the SDLC is hard. Deploying AI at scale is harder.

Writing agents wasn’t the main constraint. Infrastructure, governance, cost, reliability, and compliance quickly became the real bottlenecks.

Coming back to infrastructure with context

That realization pulled me back toward infrastructure. I reconnected with longtime colleagues from Metacloud and began working on the infrastructure problem directly, informed by what I’d just spent the last 12 months learning.

We’ve made solid early progress, and I’m looking forward to sharing more as things take shape.

Community and learning

I’ve been reconnecting with the Cloud Native ecosystem and the CNCF cloud native AI stack, and recently became a co-organizer of the Cloud Native Tampa Bay community group. This work is still early, but it’s something I’m excited to invest in next year. Alongside that, I continue to engage with the local startup community as a mentor, advisor, and peer.

In the coming year, I’ll be sharing more of this work through open source releases and educational content here and on YouTube.

Thank you to everyone who’s been part of the journey so far. Wishing you and your families a healthy and fulfilling year ahead.