Software architecture and technical leadership for teams that need to move fast.

Three things, depending on where you are.

If your system is the problem: slow APIs, databases that don't scale, infrastructure that requires heroics to keep running. I find where the drag is and fix it, at the architecture level, not just by tuning parameters.

If your team is the problem: decisions that stall, technical debt that nobody owns, engineers who spend more time firefighting than building. I bring the senior engineering judgment that helps teams move with confidence again.

If you're building something new: I help you make the right technical bets early, so you're not rewriting the foundation six months after launch. I've built products from scratch and run them in production. That's a different kind of input than pure consulting.

I've been building production software since 1998. As an employee at Sun Microsystems, as a consultant at a global digital agency, and for the past eleven years as co-founder of SoftInstigate, the company I started with a partner in 2014 to build open-source infrastructure for MongoDB and AI-powered products. The work has always been split between building our own software and helping other teams build theirs.

The architectural decisions I propose are the ones I then live with myself. RESTHeart, the open-source backend framework I created in 2014, is still running in production at companies across Europe and North America. That's a useful kind of accountability.

I write code every day. Knowing which technical bets to make, and which to avoid, is a judgment that only comes from having been wrong a few times at scale.

The code is public. You can read it on GitHub.

Architecture reviews for teams that sense something is wrong but can't pinpoint it. Technical leadership for product teams that need a senior engineering voice without a full-time hire. Hands-on work on backend systems — Java, MongoDB, AWS, API design — when that's what the situation calls for.

I tend to stay involved rather than hand off a document and disappear. The problems worth solving rarely fit a two-week engagement.

AI is not a product you buy and install. It is a layer you build on top of what you already have: your data, your processes, your existing systems. The companies that get this right treat AI as an infrastructure problem before treating it as a business problem.

In practice this means specialized agents that do specific things well, connected to real data sources, with proper governance around access, compliance and auditability. The backend has to be ready for this. Most aren't. That's usually where I start.

AI-assisted development is now part of my daily practice. It makes me faster. But that speed rests on accumulated judgment about design, architecture, testing and deployment. AI works well for me because I give it a rich, coordinated context to work in. Without that foundation, it produces fast results that don't hold up.

A cross-section of engagements across two decades and several industries:

Get in touch and describe your situation. If there's a fit, we'll figure it out quickly. If there isn't, I'll tell you that too.