Why RoboForce

The AI Residency Program

Research Focus Areas

• Vision-Language-Action (VLA) models for general-purpose robotic behavior
• World Models for predictive modeling, planning, and long-horizon decision-making
• World Action Models for jointly modeling action and environment dynamics
• Simulation and sim-to-real transfer for scalable training, evaluation, and data generation
• Reinforcement learning, imitation learning, and policy optimization for embodied agents
• Multimodal learning across vision, language, proprioception, force, and action
• Learning systems for manipulation and real-world embodied interaction

What You’ll Do

• Conduct research and build systems for embodied physical intelligence
• Develop and evaluate methods in VLA, World Models, World Action Models, simulation, and RL
• Design and run experiments on robotics tasks involving perception, planning, control, and long-horizon behavior
• Build training and evaluation pipelines for large-scale embodied learning systems
• Work closely with research and engineering teams to move ideas from prototype to real or simulated robot platforms
• Explore how multimodal foundation models can improve robot capability in real deployment settings
• Contribute to technical reports, internal research discussions, and, where appropriate, publications

Basic Qualifications

• PhD student, recent graduate, or early-career researcher/engineer in Computer Science, Robotics, Machine Learning, Electrical Engineering, or a related field
• Experience with modern ML frameworks such as PyTorch, JAX, or TensorFlow
• Experience using AI-assisted coding tools and agentic development workflows to prototype, iterate, and build quickly
• Ability to implement, debug, and evaluate research ideas in a fast-moving environment
• Strong engineering judgment, including the ability to validate, refine, and productionize AI-assisted code

Preferred Qualifications

• Rich hands-on experience in robotic manipulation, mobile manipulation, or industrial robotics
• Experience training, fine-tuning, or evaluating multimodal or embodied models
• Experience with World Models, action-conditioned prediction, model-based learning, planning, or control
• Strong hands-on experience with simulation platforms such as Isaac Gym, Isaac Sim, MuJoCo, ManiSkill, Habitat, or similar systems
• Experience with reinforcement learning, imitation learning, or post-training for robotic policies
• Experience working with real robot hardware, data collection systems, evaluation workflows, or deployment pipelines
• Demonstrated technical initiative through research, open-source contributions, or high-impact engineering work

Compensation and Resources

• Company-provided lunch and dinner, a fully stocked kitchen, and team events
• Premium fitness center membership covered by the company