Software Engineer

Role Overview

We are seeking a Research Engineer with deep expertise in state estimation, sensor fusion, and visual odometry to build next-generation perception systems for humanoid robots. This role is ideal for researchers with a strong foundation in both classical and machine learning-based estimation techniques, who are eager to translate that expertise into real-world robotic applications. You will design, implement, and deploy algorithms that enable robots to accurately estimate their motion and environment using proprioceptive and visual sensors, ensuring robust and reliable operation on hardware. Working closely with robotics, simulation, and AI teams, you will bring advanced state estimation techniques from research into high-performance, on-hardware applications.

Responsibilities

• Develop and optimise state estimation algorithms for humanoid robots using visual-inertial odometry (VIO) and sensor fusion techniques.
• Integrate classical filtering and optimization-based approaches with modern machine learning methods, including sequence-to-sequence models, autoregressive models, and diffusion-based approaches for motion estimation.
• Implement and deploy algorithms in real-time on robotic hardware, ensuring robustness and efficiency.
• Train and fine-tune deep learning models in PyTorch for end-to-end motion and odometry estimation.
• Collaborate with cross-functional teams (robotics, control, simulation) to integrate state estimation into full-stack robot systems.
• Maintain documentation, perform code reviews, and support debugging during deployment cycles.

Qualifications

• PhD or Master's degree in computer vision, robotics, or a closely related field.
• Multi-year experience (PhD or industry) in visual-inertial odometry, motion estimation, or sensor fusion for robotic systems.
• Strong C++ programming skills with experience in real-time deployment.
• Proficiency in Python and PyTorch, with experience training deep neural networks.
• Hands-on experience with robotic platforms, AR/VR devices, or similar sensor-equipped systems.
• Strong understanding of modern machine learning techniques, especially sequence-to-sequence, autoregressive, and diffusion models.
• Ability to thrive in a fast-moving, collaborative environment and take ownership of technical projects.

Benefits

• Opportunity to join a fast-moving, early-stage robotics team where every engineer has meaningful ownership and impact.
• Work on cutting-edge humanoid and manipulation technologies with significant technical depth.
• Dynamic, collaborative environment with quick iteration cycles and high autonomy.
• Competitive compensation package, including base salary and role-dependent benefits.

Location:
Zurich, Switzerland

If you're passionate about advancing state estimation for humanoid robots and bridging the gap between classical and machine learning approaches, we'd love to hear from you.