Research Engineer/Scientist - Dexterous Manipulation

Moonfire; Zürich; On-site; CHF 140,000 - 210,000; Full time

Benefits offered by this job

Competitive compensation

Enhanced pension plan

Enhanced holiday & paid leave perks

Relocation & permit sponsorship

Central Zürich office facilities

Job summary

Flexion is building the intelligence layer for next-gen humanoid robots in Zürich. We seek a Research Engineer for Dexterous Manipulation to lead VLA model development, bridging egocentric video observations with high-DOF motor control.

You will own model architecture, integrate egocentric priors with real-time policies, and ensure robust manipulation in unstructured environments. The role focuses on scalable pre-training, foundation modeling, and generative policy design, with hands-on work on

Qualifications

• PhD or Master's in Robotics, Machine Learning, or closely related field required.
• Strong background in data-driven manipulation, egocentric vision, or foundation models.
• Experience with Humanoid or Dexterous Manipulation and contact-rich physics.
• Proficient in Python and PyTorch; capable of distributed training of large-scale models.
• Proven expertise in Diffusion Models, Flow Matching, and Transformers.
• Hands-on deployment of learning-based controllers on real robot hardware.
• Experience with RL and simulation environments (e.g., IsaacLab, MuJoCo).

Responsibilities

• Architect and implement large-scale pre-training objectives for egocentric video datasets.
• Develop multi-modal Foundation Models that fuse visual perception and language into actionable trajectories.
• Design and optimize generative action heads using Diffusion Models and Flow Matching.
• Develop algorithms to align human video representations with 20+ DoF humanoid kinematics.
• Utilize Offline RL and high-fidelity simulation for policy fine-tuning.
• Translate cutting-edge scaling laws and world models into production-ready architectures.

Skills

Python

PyTorch

Diffusion Models

Flow Matching

Transformers

Reinforcement Learning

Humanoid/Dexterous Manipulation

Distributed Training

Education

PhD or Master's in Robotics/ML or related

Tools

FSDP

NCCL

IsaacLab

MuJoCo

Job description

About Flexion

At Flexion, we're building the intelligence layer powering the next generation of humanoid robots. Our mission is to accelerate the transition from fragile prototypes to real-world humanoid deployment. We are founded by leading scientists in robot reinforcement learning (ex-Nvidia, ex-ETH Zürich), and backed by leading international VC firms. In just months, we've gone from our first line of code to deploying real humanoid capabilities.

The role

We are seeking an expert in dexterous manipulation and large-scale modeling to lead the development of our physical foundation models. The goal of this position is to leverage internet-scale egocentric video to build Vision-Language-Action (VLA) models that enable our humanoid robots to interact with the world with human-like fluidity. You bring a deep understanding of how to bridge the gap between observing human actions in video and executing high-DOF (20+) motor control.

As a Research Engineer for Dexterous Manipulation at Flexion, you'll work in our Zürich office to develop and deploy state-of-the-art learning-based controllers. You will take ownership of the model architecture, integrating egocentric priors with real-time robot policies, to ensure our hardware can manipulate objects reliably and flexibly in unstructured environments.

Key Responsibilities

• Scalable Egocentric Pre-training: Architect and implement large-scale pre-training objectives for egocentric video datasets to learn generalizable representations of hand-object interactions and spatial-temporal dynamics
• VLA Foundation Modeling: Develop and scale multi-modal Foundation Models that unify visual perception and natural language instructions into actionable robotic trajectories
• Generative Policy Design: Design and optimize generative action heads using Diffusion Models and Flow-matching techniques to capture the multi-modal distribution of complex human movements
• Humanoid Motion Alignment: Develop novel algorithms to align human-centric video representations with the kinematic constraints of 20+ DoF humanoid systems, ensuring fluid and stable execution
• Reinforcement Learning & Fine-tuning: Utilize Offline RL and high-fidelity simulation fine-tuning to optimize foundation model performance for high-success-rate physical manipulation
• Cross-Functional Research: Translate cutting-edge research in scaling laws and world models into production-ready architectures that enhance robot reliability and autonomy

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 the distributed training of large-scale neural networks (FSDP, 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 (e.g., IsaacLab, MuJoCo)

Benefits

• Competitive compensation
• Enhanced pension plan
• Enhanced holiday & paid leave perks
• Relocation & permit sponsorship
• Central Zürich office with top-tier robotics testing facilities and infrastructure
• Joining Europe's leading robotics team & exposure to never-done-before research
• Energetic, collaborative culture with a bias for action and regular community events