About the role
Localisation answers one question many times a second: where exactly is the vehicle, and how sure are we? You will build the estimator that fuses lidar scan matching, GNSS with RTK corrections, wheel odometry and an IMU into a pose accurate to a few centimetres, and the map layers it localises against.
Everything downstream, from lane-level planning to hazard prediction, assumes the pose is right. Urban canyons, tunnels, snow-covered markings and construction zones all try to prove otherwise. The same estimation techniques apply to other machines that move through Clexerl ventures, so what you build is designed to be sensor-agnostic and independently testable.
The team owns its data end to end: mapping drives, map building and versioning, and online health monitoring. A localisation change is judged by replaying months of fleet logs, including the ones where it failed.
What you’ll do
- Build and maintain a pose estimator combining scan-to-map registration, inertial integration and GNSS using factor graphs or error-state filters.
- Develop the offline pipeline that builds, aligns and versions high-definition maps from fleet drives, including loop closure and drift correction.
- Detect and handle localisation degradation, such as geometric degeneracy in corridors or GNSS multipath, and report an honest covariance.
- Design map update strategies so changed road layouts reach the fleet without silently invalidating older data.
- Profile and optimise registration code to run within a fixed compute budget alongside the rest of the stack.
- Create replay tools and metrics that make localisation regressions obvious before they reach a vehicle.
What you’ll bring
- Experience building SLAM, visual or lidar odometry, or map-based localisation for robots or vehicles in production or field trials.
- A firm grasp of rigid-body kinematics, SE(3) Lie group operations, IMU preintegration and the structure of sparse nonlinear least squares.
- Proficiency in modern C++ with libraries such as Eigen, Ceres or GTSAM, and in Python for evaluation.
- Practical knowledge of point cloud registration methods such as ICP variants and NDT, and their failure modes.
- An understanding of GNSS error sources, RTK and the role of integrity in position solutions.
- A habit of validating estimators against ground truth and consistency tests, not only against visual plausibility.
Nice to have
- Experience with learned place recognition or global relocalisation after a lost-pose event.
- Work on lane-level map formats and semantic map layers that planners can query.
- Experience with long-term map maintenance, change detection and crowd-sourced map updates.
- Background in survey-grade positioning or total-station ground truth collection.
Work you might take on
- 01
Add a degeneracy detector to scan matching so the estimator down-weights lidar along featureless tunnel axes and leans on odometry instead.
- 02
Build a nightly pipeline that merges new fleet drives into the map, flags changed geometry and produces a diff for human review.
- 03
Replace a hand-tuned GNSS outlier threshold with a robust factor and show better pose continuity under urban multipath.
- 04
Create a localisation health score with calibrated covariance that the safety monitor can use to request a minimal-risk manoeuvre.
Compensation
$200K – $330K
The US base salary range for this role. Every offer also includes equity in Clexerl and, when it spins out, in the venture you help build. We tell you the reasoning behind your number.
Benefits
Health, fully covered
Medical, dental and vision for you and your dependants, premiums paid in full.
Two kinds of equity
Equity in Clexerl, plus a stake in the venture you help build when it spins out.
22 weeks parental leave
Fully paid, for every parent, however your family arrives.
Rest that sticks
Flexible time off with a four-week minimum we actually check on, plus a company-wide week in December.
Learning and compute
A yearly budget for courses, conferences and books, and compute credits for side projects.
401(k) with a 4% match
Matched from your first paycheque and vested immediately.
How we hire
Apply 5 minutes
Send links to what you have built and a few lines on why this role. No cover letter needed.
Intro call 30 minutes
A conversation with the hiring manager about the work, the team and what you want next.
Work sample 2–4 hours
A paid take-home or a pairing session on a real-shaped problem. You pick the format.
Studio day One day
Meet the team, go deep on your work and ours. Remote or in person.
Offer Within 5 days
A written offer with the full range, the equity and the reasoning behind both.
Opens an email with a short template. Links beat résumés.