Research Engineer for Dexterous Manipulation at Flexion

Location: Zurich, Zurich
Posted: Apr 21, 2026

Description

Flexion is looking for a Research Engineer specializing in dexterous manipulation and large-scale modeling to lead the development of Vision-Language-Action (VLA) models. This role involves leveraging internet-scale egocentric video to create models enabling humanoid robots to interact with the world like humans. The engineer will architect scalable pre-training objectives, develop multi-modal Foundation Models, design generative policy heads, align human motion with humanoid systems, and utilize offline RL for optimization.

Required Skills

Python
PyTorch
distributed training of large-scale neural networks (FSDP, NCCL)
Diffusion Models
Flow Matching
Transformers
Reinforcement Learning
simulation environments (e.g., IsaacLab, MuJoCo)

Additional Information

Required Previous Experiences: Experience with Humanoid or Dexterous Manipulation, including a deep understanding of contact-rich physics.
Hands-on experience deploying learning-based controllers on real robot hardware.
Experience with Reinforcement Learning and simulation environments (e.g., IsaacLab, MuJoCo)
Preferred Qualifications: Excellent knowledge of Python, PyTorch, and the distributed training of large-scale neural networks (FSDP, NCCL).
Proven expertise in Diffusion Models, Flow Matching, and Transformers.

About Flexion

Flexion

humanoid robots

Flexion Robotics is building the autonomy stack for humanoid robots, covering command to control and manipulation to locomotion across hardware and tasks. Its software uses simulation and reinforcement learning to scale to the real world with minimal human involvement. The company describes its work as building the robots’ brain and the intelligence powering the next generation of humanoids. The website presents the company as building its core team; its listed roles include an internship and engineering positions across AI and hardware.

View full company profile