Software engineer - Product industrialisation & deployment

About Verraspace

As the need for in-orbit operations grows, driven by defence requirements and the rapid increase in orbital traffic, understanding what an object is doing and predicting what it will do next sits behind most operational decisions in space. And we're doing just that.

Verraspace is a Physics AI company. We build on advanced physical models developed by our founders during their PhDs, providing a better understanding of how photons interact with their environment. The same physics that lets us read an observation more faithfully, through better image processing and object characterisation, also lets us predict orbits and trajectories more accurately. We have demonstrated both.

Physics AI is how we get from here to space world models. Physics gives structure, and holds where data is scarce; learning gives reach, and captures what the equations leave out. Combined, they become a consistent set of models, broad enough to support design-space exploration for spacecraft, payloads, and mission optimisation end to end. We intend to reach designs and mission profiles that current tools cannot, with better system performance, better mission economics, and richer, more actionable data.

The role

You will turn our models and prototypes into products customers can depend on, and build the software foundations the company will grow on. You will work directly with the physicists and ML engineers who write the models.

Where we are. Our models exist as research code and prototypes: mostly Python/JAX, developed and run on AWS (EC2). They work. They are not yet products a customer can depend on, but we have laid out an initial architecture, infrastructure and practices to support the range of applications we are about to build.

Where we are going, over the next 12 to 18 months:

• A coherent codebase, with components shared across applications and clear boundaries between models, product logic and infrastructure, without accumulating complexity we do not need.
• First production deployments with customers, including government and defence customers who run software on their own infrastructure, sometimes isolated from the internet. We need to design for those constraints from the start.
• Development infrastructure that lets a team of 15 to 20 ship quickly without breaking the models: CI/CD, reproducible environments, testing of numerical code, GPU resources for development and training.
• Performance work on the models themselves, alongside the people who write them: finding bottlenecks, optimising, and contributing features.
• Further out: running parts of our software on board satellites and maintaining it through their mission lifetime. This is not a near-term objective, but if it interests you, it is on the horizon.

What you would own. Depending on your expertise and interests, one or more of: software architecture and reusability; performance and quality of the codebase; development and deployment infrastructure. You will set the coding and architectural practices, make the technology choices and define the roadmap for your scope, with the team and on the basis of facts.

This is a foundational role, and a hands-on one. You will decide how we build software, and you will build it. The balance between architecture, infrastructure and implementation will follow the needs of the team and your strengths. As the team grows, leading the software effort is open to you if you want it; it is not a condition of the role.

How we work

We are building a company where strong engineers and scientists can do their best work. We focus on the people around you and how decisions get made.

• The mission comes first. Everyone here works on the same problem. We judge decisions on facts and on whether they make the whole system better, not on who made them.
• High expectations, in both directions. You can expect a lot from your colleagues, and they will expect a lot from you. When something does not work, whether a design, a process or a behaviour, we say so early and directly, and we fix it. We do not let problems with people linger: it is unfair to everyone else.
• Freedom with accountability. You own your scope, set the objectives you commit to, and are trusted to deliver them. Autonomy grows with the responsibility you take on.
• Trusted for judgement, challenged on substance. In your area, your call carries weight. Outside it, you learn from people whose call does.
• We solve problems, we do not create them. When something breaks, we look for the cause and the fix, not for someone to blame.
• No decoration. We do not try to make the job attractive with things that have nothing to do with the work. If what draws you is the problem and the people, you will be at home here.

A curious, versatile software engineer with real depth in at least one of the areas above and enough breadth to design a coherent architecture across them. You will become the go-to person for your domain while learning enough of the others to make sound decisions and help where needed.

• You have built from scratch. You have taken a project from 0 to 1: decided how development should work, laid down the first architecture and written the first building blocks.
• You have owned things end to end. You have worked across several stages of a software project, from early design to deployment and maintenance. You know the trade-offs and make pragmatic calls.
• You work in the code. You are comfortable analysing problems, debugging, optimising performance, developing features and refactoring architectures yourself.
• You can turn research code into a product. You have made scientific or numerical code maintainable and fast without losing what made it correct, or you want to learn how. Array computing, GPUs and reproducibility are familiar ground or ground you want to cover.
• You use AI in your workflow. You have experience with AI-assisted and agentic development, or want to build it, and can judge when these tools improve speed and quality and when they do not.

Our current stack: Python/JAX, C++, GitHub, Docker, AWS (EC2), and Kubernetes where it earns its place.

You do not need to master every area, but you should be willing and able to learn beyond your core expertise. Scope matters more than experience: we care about what you can take responsibility for, not the number of years on your CV.

The people you work with and the problems you work on are described above. We think they matter more than what follows.

• Location. Central London, near Tottenham Court Road. We work on site: the problems we solve need people in the same room.
• Relocation to Boulder. Verraspace will move to Boulder, Colorado, in early 2027, and we expect the team to move with it. We sponsor and pay for the employee's visa; we do not currently cover dependants. Keeping a London hub after the move is something we are actively exploring. Tell us your constraints in the first conversation and we will be equally direct about ours.
• Compensation. £60k to £100k+ depending on the scope you take on and your location, plus equity. Hands-on and entry-level positions start around £60k; people who take on broad responsibility go significantly higher.
• Package. A standard package for your country of employment (UK now, US after the move): pension, health cover, holiday.
• Right to work. We sponsor UK Skilled Worker visas.
• How we assess. On what you can do and how you work with others.

Hiring process.

Three steps, two to three weeks in total. You get a decision within a week of the last step and feedback at every stage, whether or not we go further.

1. Intro call, 30 minutes. You ask us anything you need to know about Verraspace; we go through the practical constraints on both sides. No surprises later.
2. Technical deep-dive, 60 minutes. We set a technical topic a week in advance. Walk us through how you would approach it, the solutions you considered, and why.
3. Conversation with the founders. The ambition, how we work, your scope and compensation.

To apply, send a CV or a link to your work and a few lines on what you would want to own here.