You’ll build the autonomy brain for a cutting-edge autonomous aerial platform that will actually take flight. At Helsing, you won’t just be developing software; you’ll be integrating state-of-the-art reinforcement learning agents into the operational systems of our own Unmanned Combat Aerial Vehicle (UCAV), the CA-1 Europa, part of the groundbreaking Centaur project. This is a unique opportunity to directly contribute to a novel autonomous system designed from the ground up.
Working at the intersection of machine learning and systems engineering, you’ll integrate reinforcement learning agents into high-performance runtime systems, enabling real-time autonomous decision-making in flight. This isn’t theoretical; your code will enable the CA-1 Europa to perceive, reason, and act autonomously in the most demanding environments.
What we build ultimately ends up in the hands of real people in high-risk, high-stress situations, so it must be both reliable and frictionless. To give some examples:
•
Autonomous Decision-Making Systems — reliable pipelines from sensor data to RL inference to tactical execution, including edge-case and failure-mode handling.
•
Reinforcement Learning Integration — bridging Python-based RL agents with Rust runtime systems for low-latency, reproducible inference.
•
Distributed Systems & Communications — handling intermittent connectivity and bespoke hardware protocols.
•
Training Infrastructure — distributed training, evaluation pipelines, and large-scale runs on custom simulators.
In some areas, we’re working at the state-of-the-art—actively implementing research papers and pushing further. In others, we’re applying proven techniques to real-world situations they’ve never encountered before. Both require skill, diligence, and deep technical understanding.
Our software operates under significant constraints, in constantly-changing environments, for users in high-risk situations. It must be reliable and frictionless. That’s what makes this work hard—and worth doing.