Robotics Engineer, Manipulation

Auxo Talent; San Francisco (CA); On-site; USD 150,000 - 230,000; Full time

Job summary

Auxo Talent in San Francisco is hiring a Manipulation Engineer to build and deploy a real-world manipulation stack for humanoid robotics. You will own end-to-end policies, from data collection to running on physical robots used in the field, and report directly to the founders while collaborating with a larger team.

The role requires hands-on experience with real robot hardware, strong Python and C++ skills for on-robot code, and a track record of bringing ideas to deployed systems with

Qualifications

• Minimum 1 year of professional experience deploying manipulation systems on real robot hardware.
• Strong Python, with C++ for anything running on the robot.
• Practical experience with imitation learning, behaviour cloning or diffusion policies.
• Evidence of taking something from a working idea to a deployed system, and honesty about what broke.
• Based in the San Francisco Bay Area or willing to relocate; on site.

Responsibilities

• Building and training manipulation policies using imitation learning and behavioural cloning.
• Owning teleoperation and data collection infrastructure, including the pipeline that feeds your models.
• Taking policies onto hardware, diagnosing why they fail, and closing the gap between testing and real-world performance.
• Working across grasp planning, motion planning and contact-rich tasks.
• Reporting directly to the founders in while collaborating with a larger team.

Skills

Python

C++

Imitation learning

Behavioral cloning

Diffusion policies

Real-robot deployment

Job description

San Francisco | On site | $150,000 to $230,000 base

Most manipulation work never leaves the lab. This one does.

We are working with a humanoid robotics company in San Francisco who need a Manipulation Engineer to build out their grasping and dexterous manipulation stack. The platform has to work in the real world rather than in a demo video. You would own policies end to end, from data collection through to something running on a physical robot that people rely on.

What you would be doing

• Building and training manipulation policies using imitation learning and behavioural cloning
• Owning teleoperation and data collection infrastructure, including the pipeline that feeds your models
• Taking policies onto hardware, diagnosing why they fail, and closing the gap between what works in testing and what works in the world
• Working across grasp planning, motion planning and contact-rich tasks
• Reporting directly to the founders in while collaborating with a larger team

What we need to see

• Minimum 1 year of professional experience deploying manipulation systems on real robot hardware. This is the one thing there is no movement on
• Strong Python, with C++ for anything running on the robot
• Practical experience with imitation learning, behaviour cloning or diffusion policies
• Evidence of taking something from a working idea to a deployed system, and being honest about what broke along the way
• Based in the San Francisco Bay Area or willing to relocate. This role is on site
• Must be a US citizen or green card holder. This client is not able to sponsor or transfer visas

Why this one is different

Plenty of labs in the Bay Area are producing impressive demos. This team is solving problems. The research translates directly into deployment, which means the work you do actually gets used rather than filmed.