Historical snapshotentry (0–2 yrs)Captured Oct 5, 2026, 11:09 a.m.

Robotics Engineer Intern

PathOn AI

Location
San Francisco, CA, United States · Onsite
Compensation
Not stated in this snapshot
Employment
Contract
Experience level
ENTRY (0–2 years exp)
Specialization
Mechatronics Engineer, Wellfound
Source
Wellfound

About PathOn Robotics

We are a pre-seed robotics startup in the San Francisco Bay Area (with a team in Shenzhen), building the application layer for robotics: one software stack, any robot, any sensor. The stack already runs six different robots and three pieces of our own open-source hardware, and it is deployed with real design partners — a factory pilot on a production line, and a school robotics curriculum going live this fall.

You would be joining a small team where a co-op ships things that reach real robots and real users, not a side project that gets archived.

The role

We are looking for a full-time, on-site Co-op Robotics Software Engineer to work directly with the founder, building and shipping robotics features end to end — from simulation through to deployment on real hardware.

This is a hands-on, high-ownership role. You will research open-source solutions, integrate them into our system, prototype quickly, and own pieces of the product. It suits someone who moves fast, learns independently, and would rather debug a real robot than tune a benchmark.

  • Minimum commitment: 6 months. Longer is preferred.
  • On-site in the Bay Area. We rent a hacker house that is both our office and where the team lives — your room there is free for the length of the co-op, so Bay Area rent is not your problem.
  • No visa sponsorship available. You must already be authorized to work in the U.S.

What you will work on

  • Build and improve grasping, manipulation, and navigation pipelines in simulation — MuJoCo, ROS2, MoveIt, Nav2, Gazebo or Isaac
  • Integrate and test ML models for robot control: GraspNet-style vision grasping, imitation learning (ACT / π0), VLA models, RL locomotion policies
  • Deploy to real robots — calibration, control, sim-to-real, and the debugging that only happens on hardware
  • Set up and maintain Docker-based environments for simulation, training, and inference across edge devices and cloud
  • Contribute to the platform infrastructure that makes all of the above reproducible for other developers
  • Research open-source robotics tools and adapt them into our stack

Requirements

  • Authorized to work in the U.S. (no sponsorship available)

  • Currently pursuing a Bachelor's or Master's in Robotics, Computer Science, Electrical or Mechanical Engineering, or similar

  • Available full-time and on-site for at least 6 months

  • Strong Python

  • Comfortable with Linux, Docker, and GPU development workflows

  • Familiar with at least one of:

    • ROS2, MuJoCo, Gazebo, Isaac, or a comparable robotics framework
    • Robot manipulation — grasping, motion planning, IK/FK
    • Imitation learning or VLA models
    • RL for locomotion
  • Able to work independently and ship with limited hand-holding

Nice to have

  • Mechanical design — CAD, 3D printing, designing a part and actually building it. We design our own hardware and this is the skill we find least often.
  • Experience with LeRobot, MoveIt, or dora-rs
  • Hardware integration: cameras, depth sensors, IMUs, servo buses
  • Full-stack ability — Next.js, Supabase, WebRTC
  • Prior startup, research lab, or open-source contribution

Why join

  • Work directly with the founder, with real impact from week one
  • Touch the entire robotics stack — simulation, ML, hardware, and platform — rather than one narrow slice
  • Your work ships to real deployments, including a factory pilot and a school curriculum
  • Free housing in the team hacker house, so the co-op is not financially painful
  • High ownership, fast iteration, and a short path from idea to a robot moving

How we hire

A short intro call, then a take-home project you can do at your own pace, then a technical conversation walking through what you built. We tell you where you stand quickly either way.