The opportunity
We’re a team of high-output generalists where ML and systems engineering converge. This is not a "run the models" role.
What you'll do
Diagnose why perception models underperform on the long tail, and turn that: into targeted data and training priorities.
Design eval metrics and regression detection that tell us whether a model is ready.
Curate and clean training data for segmentation and occupancy; hunt the data: problems that silently cap performance.
Run controlled ML experiments and ablations; cleanly separate data problems from model problems.
Debug the full data → train → eval → autonomy path, across ML and infra boundaries.
BS in Computer Science, Robotics, Statistics, Physics, Math or another quantitative area.
What they're looking for
- + years of industry software engineering experience with Python fluency and: C/C++ familiarity. Proven ability to lead cross-functional technical projects from design to completion.
- You've implemented core ML from first principles (not only called framework: APIs), and can reason about why a model learns what it does.
- Hands-on perception experience with detection, segmentation, occupancy, 3D, or sensor data.
- You're interested in solving data problems.
