About Mbodi
Mbodi is building embodied AI platform that makes robots learn and operate like humans, with natural language. Our software lets anyone teach robots new skills by talking to them and execute the learned skills reliably in production, in minutes. We are pioneering the next wave of robotics, where advanced generative models, agentic systems, and real world automation come together.
We are backed by top investors, part of YC X25, and working with global industrial partners including ABB and Fortune 100 customers across manufacturing, logistics, and labs.
Our founders bring experience from Google and robotics research at UPenn GRASP, and we are building a small, senior team focused on one of the hardest problems in AI: bringing intelligence into the physical world.
As one of our early engineers, you’ll be responsible for developing and deploying cutting-edge ML models and agentic AI systems that power robot learning and behavior. As a founding team member, you’ll play a critical role in shaping our technology, culture, and mission from the ground up.
In this role, you will:
Conduct applied research at the intersection of generative AI and robotics, with a focus on models and systems that work in the real world.
Design and implement algorithms and models for robot learning, perception, planning, and skill generalization across complex tasks.
Build robust, scalable systems that connect foundation models to reliable physical world execution, turning theory into impact.
Collaborate across ML, robotics, and product to advance agent orchestration and intelligent system integration.
Drive technical direction, architecture decisions, and research strategy from day one.
You'll thrive in this role if you:
Have strong experience in AI, robotics, machine learning, or agentic systems.
Are highly proficient in Python and PyTorch, with experience training, evaluating, and deploying neural networks.
Understand modern ML techniques such as transformers, diffusion models, vision language models, imitation learning, reinforcement learning, or robotic foundation models.
Have built and shipped real systems, not just trained models in isolation.
Are comfortable working across models, data pipelines, APIs, evaluation workflows, and runtime components.
Care deeply about reliability, latency, observability, and correctness in real world systems.
Think in terms of failure modes, edge cases, and how systems behave under pressure.
Are adaptable, execution focused, and comfortable with the ambiguity of startup life.
Why join us
This is a chance to work on one of the hardest and most important problems in AI: bringing intelligence into the physical world.
You will work directly with the founders and help shape the core platform, from agent architecture and robot learning to production deployment with global industrial customers.
We are building a small, senior team that wants to solve hard problems with real world stakes.