Research Engineer/Scientist - Dexterous Manipulation

Flexion Robotics • San Francisco, CA

2d • Onsite

About The Position

Flexion is building the intelligence layer for humanoid robots, with the goal of moving from fragile prototypes to real-world deployment. The company was founded by scientists in robot reinforcement learning and is backed by international venture capital firms.

The Research Engineer/Scientist will lead development of physical foundation models for dexterous manipulation. The role focuses on using internet-scale egocentric video to build Vision-Language-Action (VLA) models that help humanoid robots interact with the world fluidly. It requires bridging human-action observation in video with high-degree-of-freedom motor control.

Based in Flexion’s Zürich office, the role develops and deploys learning-based controllers, taking ownership of model architecture and integrating egocentric priors with real-time robot policies. The aim is reliable, flexible manipulation of objects in unstructured environments.

Requirements

• PhD or master’s degree in Robotics, Machine Learning, or a closely related field, with a strong focus on data-driven manipulation, egocentric vision, or foundation models.
• Experience with humanoid or dexterous manipulation, including a deep understanding of contact-rich physics.
• Excellent knowledge of Python, PyTorch, and distributed training of large-scale neural networks, including FSDP and NCCL.
• Proven expertise in diffusion models, flow matching, and transformers.
• Hands-on experience deploying learning-based controllers on real robot hardware.
• Experience with reinforcement learning and simulation environments such as IsaacLab or MuJoCo.

Responsibilities

• Architect and implement large-scale pre-training objectives for egocentric video datasets to learn generalizable representations of hand-object interactions and spatial-temporal dynamics.
• Develop and scale multimodal foundation models that unify visual perception and natural-language instructions into actionable robotic trajectories.
• Design and optimize generative action heads using diffusion models and flow-matching techniques to capture complex, multimodal human movements.
• Develop algorithms to align human-centric video representations with the kinematic constraints of humanoid systems with 20+ degrees of freedom, enabling fluid and stable execution.
• Use offline reinforcement learning and high-fidelity simulation fine-tuning to improve foundation-model performance for physical manipulation.
• Translate research in scaling laws and world models into production-ready architectures that improve robot reliability and autonomy.
• Develop and deploy learning-based controllers and integrate egocentric priors with real-time robot policies.

Benefits

• Competitive compensation.
• Enhanced pension plan.
• Enhanced holiday and paid-leave perks.
• Relocation and permit sponsorship.
• Central Zürich office with robotics testing facilities and infrastructure.
• Opportunity to join a leading European robotics team and work on novel research.
• Collaborative culture with regular community events.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior