About the role
ABOUT AGON
Agon is a sovereign defence AI company building a European synthetic battle arena. In it, autonomous systems train against adaptive adversaries at scale, and their failure modes are exposed before they ever deploy. Our mission is to enable decisive advantage in the era of agentic warfare while keeping human oversight, in defence of European democratic values.
Between them, our founders have helped build one of the defining defence-tech companies of the last decade, and started an aerospace company from scratch. Around them is a founding team hand-picked from Europe’s best engineers, each at the top of their field, and brought together to build this.
Following a $30M seed round in July 2026. We work with defence industry and government customers, and we’re scaling fast across our offices in London and Berlin.
THE ROLE
As a Member of Technical Staff in AI, you’ll build the intelligence at the heart of Agon. You’ll develop the AI that powers our arena and deploy it with partners. You’ll also help build our AI lab, which develops foundation models, methods, data and training infrastructure so that autonomous systems keep improving as conditions on the battlefield change.
We hire across three tracks, and the lines between them blur:
- Research scientists set research directions and generate and test new ideas.
- Applied scientists turn research into working systems, designing experiments and evaluations on real problems.
- Research engineers build the training, evaluation and data infrastructure that research runs on, and implement and scale new methods.
We’ll agree your track and level with you during the process. This role is based in London.
WHAT YOU’LL DO
- Train autonomous agents: develop reinforcement learning and related methods that let autonomous systems learn, compete and adapt in simulation at scale.
- Build foundation models for defence: contribute to pre- and post-training of models that generalise across platforms, missions and adversaries.
- Accelerate simulation: use surrogate modelling and other AI methods to make our simulation faster, and develop methods and tooling to validate it against the real world.
- Turn product needs into AI problems: define the problem, run rigorous experiments and ship what works into the product.
- Work with partners: help define partners’ problems and get AI working in their environments, alongside our engineering and delivery teams.
- Build the infrastructure: create training, evaluation and data pipelines that make experiments fast, reproducible and correct.
WHAT YOU’LL BRING
- Deep technical ability: a strong record in machine learning research or engineering, through publications, shipped models or systems you’ve built.
- Strong fundamentals: solid grounding in ML (optimisation, generalisation, evaluation and statistics) and in computer science.
- Scientific rigour: clear hypotheses, strong baselines, controlled experiments and honest reporting, including scepticism about your own results.
- Engineering excellence: excellent Python and modern ML frameworks, and the habit of writing code that’s fast, reproducible and correct.
- An AI-native way of working: you use AI tools as a default part of how you research and build, and you know where they fail.
- Research taste: you choose problems that matter, and know when to stop.
- Clear communication: you can explain your work to people outside your niche.
NICE TO HAVE
- Experience with reinforcement learning, multi-agent systems or imitation learning
- Experience with large-scale pre-training or post-training
- Background in robotics, autonomy, simulation or aerospace
- Experience building distributed training or evaluation infrastructure
- Interest in defence, or experience working with government or defence customers
WHAT WE OFFER
We pay at the top of the market we compete in, and everyone is an owner.
- Pay: salary at the 90th percentile of the companies we compete with, benchmarked by role, level and location.
- Equity: every permanent employee holds options. New hires receive options worth their base salary, vesting over four years with a one-year cliff.
- Cash and equity choice: trade some salary for more options, or some options for more cash.
- Pension: 5% employer and 5% employee contributions on full base salary.
- Ambitious work: help build AI for defence from the ground up, with research that reaches real deployments.
HOW WE WORK
Our values shape how we work: All as One Company, Take Extreme Ownership, Low Ego, Be Bold, and Remember Who We Serve. We prefer clear written thinking to polished slides, and we hold a high bar.
We welcome applicants from all backgrounds. If you’re excited by this role but don’t meet every requirement, we’d still like to hear from you.