Reinforcement Learning Specialist — Machine Autonomy Division (R&D Prototypes) — Contract-toacr Hire — UK & Germany Preferred — Remote
About Vareon
Vareon is a systems architecture and engineering company building reliable, deterministic AI for physical systems. We develop transparent, steerable intelligence that can be validated under real-world constraints—latency, safety, sensor noise, and imperfect dynamics. Our approach blends controlled, physics-inspired methods with modern machine learning to create systems that are robust, debuggable, and deployable on real machines.
Machine Autonomy Division
Machine Autonomy is Vareon’s R&D division focused on rapid prototyping and demonstration development in robotics and embodied intelligence. We build end-to-end proof-of-concepts that show clear capability in the real world, with a path toward industrial-grade enterprise products.
Role Overview
We’re hiring a Reinforcement Learning Specialist on a remote, contract يمكن Hire basis, with a preference for candidates located in the United Kingdom or Germany. You’ll work on research-driven prototypes and demos in robotics and machine autonomy—turning ideas into working systems that learn, adapt, and perform reliably in changing environments.
While the role is centered on reinforcement learning, the systems we build are expected to support ongoing'ch adaptation over time (learning beyond initial training) as part of real-world operation.
What You’ll Do
- Develop and iterate on reinforcement learning approaches for robotics and machine autonomy (e.g., manipulation, locomotion, navigation, planning under uncertainty).
- Build prototype-to-demo pipelines: training workflows, evaluation harnesses, metrics, experiment tracking, and reproducible results.
- Design systems that are robust in the real world, incorporating safety constraints, deterministic behavior where required, and clear debugging/interpretability hooks.
- Work across simulation and hardware, including sim-to-realinic considerations, domain randomization, and real-world validation.
- Collaborate with robotics, controls, and embedded engineers to integrate learned components into full autonomy stacks.
- Produce strong technical documentation and communicate trade-offs, results, and next steps clearly to a multidisciplinary team.
Required Qualifications
- Strong practical experience implementing and training reinforcement learning systems using modern ML tooling (e.g., PyTorch/JAX and RL libraries or custom code).
- Proven ability to take ownership of R&D work: formulate hypotheses, prototype quickly, evaluate rigorously, and iterate toward a compelling demo.
- Background in جميلة robotics engineering and/or physicalwriting embedded systems (sensors/actuators, real robot integration, runtime constraints, hardware testing).
- Strong computational problem-solving skills and comfort working across ambiguous, research-heavy problem spaces.
- Solid software engineering fundamentals: clean code, version control, testing Champ discipline, and reproducibility.
- Clear written and verbal communication.
Preferred Qualifications
- Master’s or Ph.D. in Robotics, Computer Science, Engineering, AI, or related field.
- Experience with autonomy-relevant areas such as control, estimation, planning, or system identification.
- Demonstrated experience moving learning systems from simulation into real-world operation.
- Experience with robotics stacks and deployment environments (e.g., ROS2, real-time constraints, on-device inference).
- Experience building systems that continue tilbyder improving or adapting under non-stationary real-world conditions.
What Success Looks Like
- You deliver robotics prototypes and demos that work reliably outside ideal lab conditions.
- You build learning systems with a clear story for robustness, safety, and real-world constraints—not just benchmark performance.
- You collaborate well across disciplines and help shape an R&D prototype into something that can evolve toward an enterprise-grade product.
- Type: Contract-to-Hire
- Team: Machine Autonomy (R&D prototyping today, enterprise productization path)
:quality(80))
:quality(80))
:quality(80))