Platform Engineer - AI and Internal Tooling

We want to use AI where it actually improves how we build and operate software. Code review is an obvious example, but there is much more scope around debugging, investigation, repository workflows and repetitive engineering work.

A large part of this role will be working out where those systems are genuinely useful, choosing the right models and tools for the job, and building the infrastructure around them so they become part of normal engineering workflows rather than isolated experiments.

Our systems generate terabytes of data, but getting useful information out of them still often requires knowing which database to query, dashboard to open or machine to inspect. You’ll help build a cleaner layer over that information so people across the company can answer questions without needing to understand where the underlying data happens to live.

The same applies to internal knowledge spread across code, documentation, internal messaging, incident notes and dashboards. We want that context to be easier to find and use, while still making it possible to get down to the raw source when needed.

You’ll build the tools that remove friction from how people at Gattaca work. That might mean improving CI and deployment workflows, building better interfaces around existing systems, or replacing a manual process with something simple and reliable.

Production systems generate a constant stream of information through alerts, logs, crashes and traces. Today, investigating an issue often means manually piecing together context from several places.

We want to make that loop much tighter. If a builder panics overnight, an agent should eventually be able to gather the relevant logs, inspect the host and codebase, understand what happened and prepare a fix for an engineer to review. If an alert appears in our internal messaging we should be able to investigate it from there rather than starting again from scratch.

The more useful these tools become, the more access they need. An agent that can inspect production, read source code or prepare changes is useful precisely because it can interact with real systems.

You’ll help design the boundaries around that access: what tools can see, what they can change, when a human needs to approve something and how everything is audited. The aim is to make these systems powerful without making them opaque or difficult to control.

We believe the best work comes from small, focused teams with full ownership over their domain. Deciding what to build and how to build it.

We keep hierarchy as minimal as possible. Engineers own problems, not tasks. You'll have significant autonomy from day one. After a few months you'll be driving the full loop: identifying something worth improving, understanding the problem properly, building it, deploying it and seeing whether it actually made things better.

That matters particularly for this role because many of the most useful projects will not arrive as clearly defined tickets. The tools available are changing quickly, and a large part of the job is noticing where a new capability can remove work or make information easier to access.

We deliberately keep the engineering team small and hire very selectively. That gives this role unusual leverage: a good internal tool or workflow improvement can make a meaningful difference to the output of the whole team. You’ll work alongside engineers from Jane Street, Citadel Securities, Tower Research, Flow Traders, Meta, Google, as well as people from startups and less conventional backgrounds.

High agency. A lot of the work will start as something vague. You should be comfortable finding these problems yourself, deciding what matters, and taking them from observation through to something people will actually use.

Strong software engineering. We have no preference on language. But you should be comfortable building complete systems and moving across backend services, data, infrastructure and simple user interfaces when the problem requires it.

Good product judgement. Internal tooling is only useful if it makes people's work noticeably easier. You should have a good instinct for where automation removes real friction, where a simple tool is enough, and where introducing another system would make things worse rather than better.

High drive. There is a lot to improve, and platform engineers can be huge levers on team output. We want someone who moves quickly, raises the standard around them, and keeps pushing important work forward without needing a queue of tickets.

London-based, in-person. The problems we work on are hard enough that the value of in-person collaboration is extremely high. We're very flexible when life requires it, you have agency over how you get the work done, but the default is being in the office with the team.

We sponsor UK work visas for candidates who need them.