Software Engineer, Robotics

TAR•San Francisco, CA

1d•Onsite

About The Position

As a Software Engineer on the robotics team, you own the software that makes those machines autonomous: middleware, perception, localization, planning, and control on the machine, and the fleet layer that coordinates dozens of them working the same site at once. This is a mid to senior level individual contributor role based at our San Francisco robotics facility. Your code runs on real hardware, outdoors, and you will travel to deployment sites as-needed and be there when it does.

Requirements

Professional experience building software for robotic or autonomous systems.
Track record shipping production robotics software on real, deployed hardware, not solely research or lab work.
Expertise across the autonomy stack, from fast prototypes to production code in a systems programming language such as C, C++, or Rust.
Experience building for real-time and resource-constrained environments.
Ability to take a system a long way independently, from architecture through field deployment.
High ownership, bias for speed, and comfort with ambiguity.

Nice To Haves

2+ years of professional Rust experience in a real time, safety-critical environment.
Outdoor, off-road, heavy-industry, or construction autonomy.
GNSS/RTK, SLAM, or large-scale outdoor localization.
Multi-robot coordination and fleet management.
Deep expertise in one or more of perception, state estimation and sensor fusion, motion planning, or controls; or a strong generalist who ships features across the entire stack.

Responsibilities

Write software across the autonomy stack for prototype and production robots: perception, state estimation, localization, motion planning, and controls.
Implement task and motion planning that places physical assets with centimeter tolerances at construction pace.
Build the fleet layer: task allocation, coordination, and deconfliction across multiple machines operating on an active construction site.
Stand up simulation and CI so changes are validated before they reach a machine, and build the logging, telemetry, and data pipelines to debug what happens when they do.
Build the operator-facing tooling for teleoperation, supervision, and diagnostics that field crews depend on daily.
Ship software that runs offline and degrades gracefully at remote sites with limited connectivity.
Go to the field, debug on the machine, and close the sim-to-real loop with the mechanical and electrical teams.