At Devol Robots, we're redefining how robots learn and interact with the physical world. We're a robotics AI company pioneering physics-aware embodied intelligence, built around our Interaction World Model, a physics-aware AI model that learns the physics of contact and manipulation, enabling robots to anticipate and reason about how the world responds when they touch, grasp, and move through it. Rather than relying on computer vision or motion planners alone, our models understand force, contact, and dynamics the way a task actually unfolds, so robots don't just see the world, they feel and predict it. Our proprietary technology has already been deployed to customer across Singapore, China, and the United States. Join us to shape the core AI infrastructure that will move robotics from rigid automation to adaptive, intelligent systems.

You'll work alongside our CTO and a world-class team of AI and robotics experts, they are builders at the intersection of hardware, software, and artificial intelligence, as we refine, scale, and release production-ready platforms. This is a rare opportunity to define how robots master physical interaction for the next decade and to help drive the next paradigm shift in physical AI.

Job Summary

We are seeking exceptional AI Engineers to advance our physics-aware embodied AI models. In this role, you'll develop and deploy our proprietary AI models that empower robots to perceive, reason about, and adapt to physical forces, contacts, and dynamics in real-time. We're looking for talent experienced in large-scale model training (e.g., generative models, pre-training, or post-training), deploying massive AI systems, or building applications from existing models. If you have hands-on AI development, deployment, or application experience and are passionate about creating foundational models for robotics, we welcome you to apply. This is your chance to tackle unsolved challenges in embodied intelligence at a high-impact, revolutionary robotics AI company.

Your Role

Design, develop, and scale our in-house world models using our proprietary AI representation and learning methods to push the boundaries of physical interactions in robotic systems.
Build and train large-scale AI models for robots to feel and experience physics, incorporating data from sensors (e.g., force/torque, tactile, vision) to enable robust interaction in tasks like assembly, grasping variable objects, or recovering from disturbances.
Implement advanced algorithms for physics-aware learning, including generative pre-training, post-training fine-tuning, and integration with control systems for real-world deployment.
Create and manage end-to-end ML pipelines: collect interaction datasets from physical robots, perform unsupervised pre-training on large-scale data, and deploy models on edge devices for low-latency inference.
Develop applications for customers by connecting AI models directly to our proprietary SDK and control layer, optimizing for production environments and ensuring seamless integration with robotic hardware.
Optimize models for scalability, reliability, and safety in dynamic settings, using techniques like quantization, distributed training, or federated learning to handle real-time physics simulations.
Conduct experiments, ablation studies, and validations in simulation (e.g., Issac Lab, MuJoCo) and hardware, iterating on metrics like interaction success rates, adaptability to novel scenarios, and computational efficiency.
Collaborate with robotics engineers and product teams on technical roadmaps; participate in design reviews, mentor peers, and contribute to publications in embodied AI.
Research and develop cutting-edge innovations in physics-aware embodied AI to enable lifelong adaptation and robust physical interaction in real-world robotic applications.

Qualifications

PhD or MS in Computer Science, Artificial Intelligence, Robotics, Machine Learning, or related field.
3+ years of experience in AI/ML, with a focus on large-scale model training (e.g., generative models like VAEs/GANs/Diffusion, pre-training transformers, reinforcement learning algorithms like PPO/SAC, or post-training alignment).
Proven track record in deploying large-scale AI models to production or developing applications from foundational models (e.g., adapting LLMs/VLMs for specific domains).
Strong proficiency in Python and C++ for AI development; expertise in ML frameworks (PyTorch, TensorFlow, JAX) and tools for distributed training (e.g., Horovod, DeepSpeed).
Hands-on skills in ML pipelines, including data curation for unsupervised learning, model optimization, debugging, and version control (Git).
Passion for foundational robotics AI; anyone with AI dev/deployment/application experience interested in embodied world models is encouraged to apply.

Preferred Qualifications

Experience with large scale model training, physics simulation (e.g., Issac Lab or MuJoCo for contact-rich tasks), or unsupervised learning for embodied agents.
Solid foundation in robotics and physics: dynamics, kinematics, control theory, sensor fusion, and experience with robot datasets.
Publications in top venues (e.g., NeurIPS, ICML, CoRL, ICRA, IROS, RSS) on large-scale AI, embodied AI, generative modeling, or robotics.
Background in deploying AI in real-time systems, edge computing, or industrial applications.
Familiarity with open-source contributions in AI/robotics, or experience at AI robotics companies.

Why Join Us

Take part in building a foundational technology that will set the standard for physics-based embodied AI in robotics.
Work with a world-class team of AI researchers, robotics engineers, and system builders solving unsolved challenges in physical interaction and intelligent manipulation.
See your AI models deployed in real factories, households, and production environments worldwide—powering adaptive robots for global customers.
Join a fast-growing company with deep technical roots, bold vision, and global ambitions, backed by proprietary, patent-pending technology.
Contribute to a mission that is not incremental automation, but a paradigm shift in how robots understand, feel, and master the physical world.

This is more than a job—it's a chance to pioneer the next generation of embodied AI. You'll be at the heart of a movement to revolutionize how robots perceive force, contact, and interaction—laying the foundation for intelligent systems that will transform industries worldwide.

This is a once-in-a-generation opportunity to shape how robots will experience and adapt to physical environments for decades ahead—driving the global rollout of groundbreaking technology poised to redefine the future of work.

If you're an AI engineer driven to build scalable, production-ready, and revolutionary embodied AI systems—we invite you to help shape the future with us.

Senior AI Engineer - Physical AI