Open position
Staff Software Engineer (AI Lead)
We are seeking a Staff Software Engineer to set the technical direction of the AMP AI team. The team owns two connected problems: the agentic framework that turns operator intent into executable workflows, and the perception behind vision-in-the-loop manipulation.
You will define the approach, build the team, and carry accountability for shipping the capability into customer deployments.
You will work with the AMP architect, the core robotics teams, and product, and you will own the interface between learned components and the deterministic execution path.
What you’ll do
- Set the technical direction for the AMP AI team across the agentic framework and perception.
- Own the architecture of the intent to workflow path: semantic reasoning, decomposition, and the handoff to orchestration.
- Define how learned components integrate with deterministic execution, including bounds, fallback, and validation.
- Lead the perception direction for vision-in-the-loop manipulation: palletization, depalletization, bin picking, and machine tending.
- Establish evaluation methodology that measures model quality against deployment outcomes.
- Build and mentor the AI team, including hiring and technical growth of engineers.
- Drive the data strategy with the Data Collection Lead so training data reflects real deployment conditions.
- Contribute directly to the hardest implementation problems in the stack.
Job requirements
Experience
- 8 to 12 or more years of professional engineering experience, with 5 or more years in applied machine learning or robot learning.
- Demonstrated experience shipping learned components into production robotics or another physical system with real consequences for failure.
- Experience leading a technical team or workstream, including hiring and mentorship.
Education
- Master's degree or PhD in Robotics, Computer Science, Machine Learning, Electrical Engineering, or a closely related discipline.
- Equivalent advanced industry experience with demonstrated technical leadership may be considered.
Core technical background
- Deep expertise in one or more of: robot manipulation learning, 3D perception, vision language models, or vision language action models.
- Strong grounding in classical robotics as well as learning, including grasp synthesis, pose estimation, and sensor fusion.
- Experience with the full model lifecycle: data, training infrastructure, evaluation, deployment, and monitoring on constrained compute.
- Experience with LLM-based agent architectures, tool use, and structured output, and a clear view of where they fail.
- Proficiency in Python and C++, and experience with PyTorch and modern training infrastructure.
- Hands-on experience with ROS or ROS 2 and with deploying models on robot hardware.
Systems and architectural mindset
- Ability to decide what should be learned and what should not.
- Comfortable designing systems where a model sits inside a system that has to stay predictable.
- Able to judge whether an approach that demonstrates well will hold up in a facility.
Technical leadership and collaboration
- Operates as a cross-functional technical authority on AI direction.
- Communicates the state of the capability honestly to product and executive stakeholders, including what does not work yet.
- Clear communicator who can work effectively with remote teams across multiple time zones.
Nice to have
- Experience with imitation learning or reinforcement learning on physical robots.
- Experience with large behavior models or robot foundation models.
- Experience with simulation to real transfer and domain randomization.
- Publication record in robotics or machine learning.
- Prior open-source contributions to the robotics or ML ecosystem.