Matt Jesuele

I'm the founder and principal at Ansatz — an engineering practice serving software teams and investment managers.

The Thesis

AI made production cheap. That should've been a revolution. For most teams, it wasn't.

What actually happened: teams got faster at producing code without getting better at engineering. More output, same architecture, same tribal knowledge, same fragile patterns. AI amplified whatever was already there — including the bad parts.

The bottleneck shifted. It's no longer "can we build it." It's "do we know what to build, and will it hold together at scale?"

Most teams adopted the tools without changing the process. More output, same problems, new ones on top.

That gap — between having AI tools and actually getting leverage from them — is where Ansatz operates. Someone has to bring engineering rigor to how teams build with AI. Not tools. Not hype. Judgment.

The Background

15+ years building software. Self-taught as a teenager, formalized at Columbia. TypeScript mostly, but I'll work in whatever the codebase needs.

Co-founded a technical infrastructure company and scaled it to $50k MRR. I know what systems look like when they work and when they break.

I work inside real codebases with real teams. Not advising from the outside. Production code ships every week.

I also trade my own capital and build systematic trading infrastructure. The quantitative, systems-oriented thinking isn't a consulting persona — it's the default mode.

The unconventional path is the point. The perspective comes from doing the work, not from studying it.

The Work

Two practices. For software teams: fractional senior engineering and agentic architecture — the judgment layer around how a team builds with AI. For investment managers: ops automation, reporting infrastructure, and turning investment processes into software.

One principal. Same hands on the spec, the code, and the handoff. No associates, no partners who sell the work someone else does.

Seeing many codebases and many funds is the structural advantage. A full-time hire goes deep in one. Across engagements, every project sharpens the pattern matching that makes architectural decisions good.

Ships code, not decks. The deliverable is working software, not recommendations about working software.