I come in, fix what's slowing you down, and leave.

My name is Leo Riviera. I work with founders and engineering leaders where technology has become the bottleneck. Fixed-scope project work: architecture, modernisation, internal tooling. Clear outcome agreed upfront. When I'm done, your team is stronger, your systems are more resilient, and your codebase is in better shape than I found it.

Book a call
What I do, and who it's for

Three types of project work

I do my best work with companies at a specific moment: product is working, the team is growing, and engineering has quietly become the thing that's slowing everything else down.

  • Strategy

    Architecture & Technical Strategy

    You get a clear picture of what's wrong, what to fix first, and why. I look at your codebase, your team, and your roadmap. Then I tell you what's actually blocking you, not what's easiest to say. I use systems thinking to understand how decisions in one part of the organisation create drag everywhere else. If needed, I'll lead the execution too.

    Good fit if

    Direction is unclear, or the architecture has quietly become a constraint. Decisions keep getting deferred. Typically Pre-Seed to Series B, with a team of 5 to 30 engineers.

  • Modernisation

    Codebase & Stack Modernisation

    Legacy systems create drag on hiring, velocity, and reliability. I untangle the mess, migrate to infrastructure that makes sense, and sequence the work so delivery keeps moving. Your engineers are in the room throughout. I'm not a black box.

    Good fit if

    You're hiring but onboarding takes too long. Velocity has dropped and the codebase is the reason. Series A to C, 20 to 100 engineers.

  • Effectiveness

    Engineering Effectiveness & AI Tooling

    I build the internal tools and workflows that reduce toil and help your team move faster. Less time on repetitive work, faster onboarding, sharper focus. The goal isn't just faster shipping. It's an engineering function that's self-organising and resilient: one that keeps moving well without constant oversight.

    Good fit if

    Too much engineering time goes on toil, slow reviews, or fragile manual processes. Any stage from Seed onwards, typically 10 to 60 engineers.

How every engagement works

  1. Understand

    I look at the codebase, the team, and the roadmap before I form any opinions. Every situation is different and I don't assume otherwise.

  2. Align

    We agree on what success looks like and in what order to tackle it. Near-term progress matters, but not at the cost of making things harder later.

  3. Execute

    I'm in the code and in the decisions. No slide decks. You'll see real progress, not status updates.

  4. Hand over

    When I leave, everything is documented and your engineers are briefed. The goal is a team that doesn't need me anymore. Anything less would be a poor result.

Technologies I work with most

Node.js, TypeScript, Python, Go · AWS, GCP, Terraform · PostgreSQL, Redis, MongoDB · React, Next.js · Docker, Kubernetes, CI/CD pipelines.

Not limited to the above. If your stack isn't listed, it's worth a conversation.

Common questions

Things people wonder before hiring me

Bringing in an outside engineer for a fixed project is a different decision to hiring. Here are the things that come up most often.

  • You won't get up to speed fast enough.

    I've done this enough times to be fast at it. The first few days are for listening and reading. I don't touch anything until I understand it. Most clients are surprised how quickly that happens.

  • One engineer can't change much.

    I'm not a headcount solution. Senior judgement applied at the right moment changes what a whole team can do. The bottleneck is almost never people. It's decisions that keep getting deferred.

  • What happens when the project ends?

    Everything is documented. Your engineers know what was built and why. I'm not trying to make myself indispensable. A clean handover is part of the job.

  • We just need more time, not outside help.

    Sometimes that's right. But more often the problem isn't time. Hard decisions keep getting pushed, and the backlog has been quietly growing for longer than anyone wants to admit.