Daniel Mestas. Builder, manager, operator.

I’m an engineering manager who codes, focused on agentic engineering and AI-native solutions. 14 years shipping software, 4 years leading teams. I spend real time in the codebase every day, and I’m usually the person the team calls when something is hard to debug.

The work I’m most interested in right now is making engineering teams ship like significantly larger ones — by putting real AI tooling in the hands of every engineer, writing the shared toolchains and workflows that make it work, and pairing with the team until the practices stick. I’ve done this end-to-end: built the infrastructure, ran the pilots, presented to leadership, changed company-wide AI policy.

On the production side, I ship AI-native features that hold up — retrieval-augmented experiences, sentiment-driven workflows, automated escalation, and agent loops that know when to stop. The thing I keep optimizing for is software that quietly works, not features that demo well.

The rest of my career has been on applications that run companies: system architecture on AWS and Azure, frontend in React, Vue, and Angular, backend in a handful of languages. Line-of-business systems, operator software, large-donor management platforms, fintech and insurance back ends — the kind of software a real team relies on every day.

I lead from inside the codebase, not from a dashboard. My management philosophy: ownership first, craftsmanship second, transparency third, pragmatism as the guardrail. I don’t micromanage when things get hard — I work harder, slow the tempo, and tighten the guardrails.

Outside the day job I build local-first infrastructure in Go — open-source utilities for autonomous coding agents (libfossil, DagNats, EdgeSync), a point-of-sale system for an independent motorcycle shop, and side projects when something interesting comes up.

This site is a home base. If you saw something here you’d like to talk about, send me an email.