The Story

I built and sold my first company at 23. I climbed to principal engineer at VC-backed companies. Then I spent 12 years inside agencies, building software for other people's businesses.

I saw every failure mode. Scope creep disguised as "agile." Junior devs billed at senior rates. Codebases nobody could maintain. Teams that were perfectly capable but had no shared system for thinking about what they were building. I watched smart people write bigger and bigger checks for the same problems, because nobody had the incentive to fix the root cause. Agencies profit from ambiguity. They don't profit from clarity.

At some point I realized the problem was never the people. It was always the system. Teams don't need more resources. They need a better way to think about what they're building. When you give them that, everything changes. Communication opens up. Decisions get easier. The work starts compounding instead of spinning.

Now I work directly with founders as an embedded engineering partner. One product. One relationship. For as long as you need.

I do not advise from the sidelines. I build.

Philosophy

Never rewrite.

The most expensive decision in software is starting over. Almost every rewrite is avoidable. The problem is rarely the code. It's the absence of a system that defines what the code should do.

Define before you build.

If nobody has written down what the product does and why, everything downstream breaks. Estimates are wrong. Quality suffers. Velocity erodes. The spec isn't bureaucracy. It's the foundation.

Your team isn't the problem. Your system is.

Fix the system and the same team performs dramatically better. I've seen it happen over and over. The people don't change. The output does.

Compounding beats rebuilding.

Every cycle should ship value, reduce debt, and raise quality. All three, every time. Software development should compound, not just accumulate.

AI is a multiplier, not a replacement.

AI exposes how good or bad your process is. Companies with documented business logic, structured specs, and clean architecture can leverage AI at a level their competitors can't touch. Without that foundation, AI just produces bad code faster.

Your product needs a partner, not a project.

The longer I work with a product, the faster everything moves. Context compounds. That is by design.

Credibility

12+ years building software. Built and sold a company. Principal-level engineering at VC-backed companies. Ran a development agency and saw every failure mode firsthand.

I built a spec-driven methodology called SpecOS from years of real engagements. It's the operating system that makes agentic development accountable, not just fast.

I know what breaks and why. That's not something you learn from a course. It's something you learn from being in the room when things go wrong.

A founder was spending $8K a month on an offshore team and still managing every detail. Within a few months of working together, he had one partner who understood the product, the codebase, and the roadmap. He stopped managing developers. He went back to running his business.

He went on vacation without his laptop for the first time in years.

He told me he would like to work with me for the rest of the time he owns the product.

Spending money on dev and not sure it is working? Let's talk.

Book a Discovery Call