AI lets teams build faster. The new challenge is making sure what we build is ready, that we can defend it, and that we understand it thoroughly.
I’m taking on the role of VP, Engineering at Elevata. We have a team of experienced engineers who know their craft. My job is to build on that and make it stronger as AI changes how fast we can build.
Building fast is no longer the hard part. DORA’s research finds that the time AI saves in writing code is “frequently re-allocated to auditing and verification”, and that AI “magnifies the strengths of high-performing organizations and the dysfunctions of struggling ones.” Experienced engineers who stand behind their work are exactly the kind of strength AI magnifies. Making AI speed dependable is how we turn that into the next engineering advantage.
My mandate covers our engineering standards and how we build and grow our engineering teams. I’ll do it alongside our CTO, Paulo Frugis, and our Chief AI Officer, Bruno Machado Valerio, following the direction set out in our recent leadership announcement.
Ready, defensible, understood
These are what make speed worth something to a customer, and they’re what I’ll keep strengthening as we go faster:
- Ready: it works end to end against what we agreed, not just that the code is written.
- Defensible: every decision in it has a reason we can explain and evidence that it holds.
- Understood: the team knows the solution thoroughly, not just the parts each person wrote.
You own what you ship
One standard sits under everything else: you own what you ship, whether you wrote it or AI did. In review, the engineer has to explain what the change does, why it’s built that way, and how they know it works. “The AI wrote it” isn’t an answer. If they can’t defend it, it doesn’t ship.
That applies to infrastructure, architecture documents and runbooks as much as to application code. It also means keeping changes small enough to review. Large AI-generated changes don’t remove risk. They push it into testing and deployment, where DORA finds AI productivity gains are often absorbed.
One failure mode deserves a name. If a model writes the code and the tests from the same wrong reading of a requirement, the tests pass and the requirement is still unmet. Tests need a basis outside the code they check: acceptance criteria, observed behavior, and an engineer who understands the problem.
Experience, amplified
AI amplifies the experience and skill already there. A senior engineer who knows what to ask, what to cut, and when an answer is wrong can now do what used to take a team. An engineer without that judgment just produces more code they can’t defend.
So I want small teams led by experienced engineers who use AI every day, stay hands-on, and treat mentoring as part of the job, so that judgment spreads.
What customers should see
You bring us a problem worth solving. We bring a small, experienced team that uses AI to move fast, and the judgment to make that speed count for you.
Bigger ideas, within reach. With AI in experienced hands, we can explore more options before committing to an architecture, take on integrations that used to be out of reach, and get ideas in front of your users sooner.
Ready when it’s delivered. You receive the solution we agreed on, tested against the success criteria we set together up front and ready for real users.
A system you truly own. Every part comes with engineers who can explain what it does, why it’s built that way, and how we know it works, and we make sure your team understands it too. It’s yours to run, change, and grow, not a black box.
Speed is the easy part now. What we build has to be ready, defensible, and thoroughly understood. That’s what making AI speed dependable means to me.
Talk to Elevata about planning and delivering your next cloud or AI project.
