Software engineer and technology lead. 15+ years in IT.

I work across backend systems, infrastructure, and EVM smart contracts, with a growing focus on AI in engineering and how it's changing the way software actually gets built. Most days I'm doing some mix of hands-on coding, system design, and leading engineering teams — the ratio depends on what the project needs at the moment.

Node.js and TypeScript for backend services. Solidity for EVM and DeFi work. React, Angular, or Vue on the frontend, depending on what the project already runs on. Postgres, Redis, message queues, observability, the usual production stack on AWS or GCP. I'm not particularly attached to any specific tool — the right stack is whatever fits the problem and the team that has to live with it for the next few years.

AI tooling is part of my daily workflow now — prototyping, refactoring, code review, debugging, research, most of the cleanup work that used to be slow. The productivity shift is real, but only if you stay honest about where AI actually helps and where it just produces confident-sounding nonsense.

What I find most interesting isn't whether AI can write the code. It's the bigger shift behind it: how teams, codebases, and engineering processes change when capable agents become normal infrastructure. Code review takes a different shape. Ownership models need rethinking. New kinds of failure show up in production. That's where most of my attention is going right now.

I prefer small, reversible changes over big rewrites, and I'd rather understand existing code than replace it. Tests are part of the work, not a separate phase. I'm slow to add abstractions and quick to remove the ones that aren't earning their cost. I aim for production-grade on the first pass because anything less just becomes someone else's problem later.

When I'm leading, the approach is pretty simple: make ownership clear, keep standards visible, cut meetings down to what's actually needed, and lean on written context for the rest. I trust people to do good work and try to keep the noise out of their way.

A lot of my exploration time right now is on agentic development workflows — figuring out where they're actually ready for production use, where they aren't yet, and what the day-to-day shape of software engineering will look like a couple of years from now.