A small lab, and the systems it ships.
Cogvera Labs is a forward-deployed AI engineering practice based in Perth, Western Australia. We’re small by choice, technical by trade, and built to embed inside the organisations we serve.
Most AI never reaches production. Getting it there is the job.
The gap between an impressive demo and a system a team relies on is enormous. Almost every AI initiative dies there, quietly. Bridging it takes someone who will sit inside the work, govern from the first line of code, and build something that keeps running under whoever ends up holding it.
We do nothing else. No platform to sell, no seats to fill: systems in production, owned by the people who use them.
Thomas Nicklin
Founder · Forward-Deployed Engineer
Thomas Nicklin is an engineer in Perth, Western Australia. He founded Cogvera Labs because the best AI opportunities inside an organisation are almost never the ones on the roadmap. They sit in the day-to-day work, visible to the people doing it and to nobody else.
Cogvera’s answer is to embed: Thomas joins the client team, opens with a 2-hour immersion, and stays for the build the immersion scopes. He writes the code himself. An agent-augmented toolchain supplies the throughput a team would otherwise supply, and a gate on every commit decides whether any of it ships. Governance goes in early, and the data plane stays onshore in Australia. When it ships, the client decides who runs it from there: their own team, ours, or both of us on terms agreed before the first commit.
Damian Mazzone
Capture Manager
Relationships are where I start every deal. My love for connecting with people has always come naturally, but it was years of hands-on project management, across public, private, and government sectors, that turned it into a real skill: capture management. I've learned that success isn't just about strategy and process, it's about listening first, understanding what people actually need, and building the kind of trust that turns a first conversation into a lasting partnership. Different sectors bring different cultures and priorities, but the approach stays the same: people first, results follow.
Ruoyu (James) Wang, MSc
Principal Systems Scientist · MD-Candidate
I've always been passionate about research and the pursuit of discoveries that can improve patient care. Working across biomedical and clinical research has reinforced my belief that meaningful innovation comes from combining scientific curiosity with practical solutions.
I see AI as the next frontier in advancing medical research. By helping researchers analyze complex data, uncover new insights, and accelerate discovery, AI has the potential to fundamentally reshape how we develop the next generation of diagnostics and treatments.
One engineer, and a lot of machinery.
Firms this size usually imply a bench. You would work out the real number soon enough, and it should change how you read the rest of this site, so here it is up front.
One engineer writes every line we ship. The throughput that would ordinarily need a team comes from an agent-augmented toolchain: code arrives quickly, then has to survive gates that do not care what produced it. Type checking, linting, dead-code analysis, tenant-isolation tests, accessibility tests and static analysis run on every commit, and nothing merges until all of them pass. Judge us on the gates rather than on the headcount, because the gates are the part you can check.
The other people on this page work on client development and domain research. They do not build the systems, and we do not describe them as though they do.
We run one build at a time. While an engagement is live, the next one waits. That is the honest capacity of this arrangement, and it is why the immersion scopes your work before anyone commits to a date.
What this costs you
- No parallel workstreams. A programme that needs several teams moving at once is not something we can staff, and we will say so rather than take it.
- No round-the-clock on-call. Support hours are agreed per engagement in writing, and they are Australian business hours unless the contract says otherwise.
- No bench to reassign. If the work needs a second engineer, you hear that before you sign.
The obvious question about a one-person firm.
What happens to your system if the person who built it is unavailable? It is the right question to ask us, and the answer is not a reassurance. It is four properties of how the work is built, each of which you can verify while it is being built.
What this is not
This is a continuity argument, not an insurance policy and not a source-code escrow arrangement. If your procurement process requires either, raise it in the first conversation and you will get a straight answer about whether we can meet it.
Four commitments.
Perth, mostly.
Headquartered in Western Australia, building for teams across the country and beyond. The data plane stays onshore; the engineering deploys forward.
Bring us a problem.
The first session is two hours with your team, on the real work.