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

As our VLA/WAM Research Engineer, you'll develop the foundation models that power our humanoid robots, spanning model architecture, large-scale pre-training, adaptation, and post-training.

You will work across the model lifecycle: designing VLA and World-Action Model architectures, training foundation models on large and diverse embodied datasets, developing methods for efficient adaptation to new tasks and embodiments, and using robot experience to continuously improve model capabilities.

In particular, we are looking for researchers and engineers with deep experience in one or more of the following areas: VLA/VLM or multimodal model architecture, large-scale foundation-model pre-training, world models and predictive learning, robot learning, or post-training and continual learning. If you have strong depth in one of these areas and are excited to work across the broader problem of building general-purpose robot foundation models, you will be a strong fit for the role.

Key responsibilities

• VLA/WAM Architecture: Design and develop multimodal foundation models combining vision, language, robot state, actions, and temporal context for general-purpose robotic control.
• Foundation Model Pretraining: Develop and execute large-scale pretraining across diverse robot embodiments, teleoperation, egocentric human data, and simulated experience.
• Training Recipes & Scaling: Improve data mixtures, training objectives, optimization strategies, and distributed training workflows to deliver increasingly capable foundation models on a regular cadence.
• World-Action Models: Develop predictive world and action models for representation learning, future-state prediction, planning, and robot policy improvement.
• Adaptation & Mid-training: Develop methods to efficiently adapt pretrained models to new tasks, environments, embodiments, and capabilities.
• Post-training & Reinforcement Learning: Improve models using real-world and simulated robot experience, including imitation learning, RL, reward/value-guided learning, and specialist-to-generalist distillation.
• Continual Learning: Develop methods to learn from autonomous rollouts, failures, interventions, and recoveries while retaining existing capabilities and avoiding catastrophic forgetting.
• Evaluation & Research: Design rigorous experiments, evaluate models in simulation and on real robots, and use results to guide architecture and training decisions.

Requirements

• Strong Python and PyTorch skills, with hands-on experience developing and training modern deep-learning models.
• Research or engineering experience in one or more of the following: VLA/VLM architectures, multimodal foundation models, world models, large-scale pretraining, robot learning, or RL/post-training.
• Experience owning substantial model development or training experiments, including architecture choices, training objectives, optimization, and evaluation.
• Solid understanding of modern deep-learning architectures, representation learning, and the relationship between data, training objectives, and model performance.
• Ability to design rigorous experiments, diagnose training failures, and use empirical results to guide research decisions.
• Familiarity with large-scale datasets, model training workflows, and the challenges of generalization across tasks, environments, or embodiments.
• Ability to critically evaluate recent research and translate promising ideas into working models and experiments.
• Comfortable working across model architecture, training, and adaptation while bringing particular expertise in at least one of these areas.

Additionally, the following are a plus but not required:

• Experience training VLA, VLM, diffusion, flow-matching, video/world models, or other multimodal generative architectures.
• Experience with distributed training, multi-node GPU systems, large-scale data mixtures, and foundation-model training optimization.
• Experience with reinforcement learning, offline RL, reward/value learning, specialist distillation, or continual learning.
• Experience with predictive world models, video generation, latent dynamics, or self-supervised representation learning.
• Experience with multi-embodiment robot learning, action representations, or adapting policies across different robot platforms.
• Experience working with teleoperation, egocentric video, simulation, autonomous rollouts, or other large-scale embodied datasets.
• Experience evaluating learned policies on real robots, particularly humanoids or dexterous manipulation systems.

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
• Competitive Compensation
• Joining a leading robotics team & exposure to never-done-before research
• Energetic, collaborative culture with a bias for action and regular community events

Zurich

• Enhanced pension plan
• Relocation & permit sponsorship
• Enhanced holiday & paid leave perks
• Central Zürich office with top-tier robotics testing facilities and infrastructure

San Franciso

• 401(k) with company contributions
• Health, dental & vision coverage with the flexibility to choose your own plan
• Open PTO policy & paid company holidays