Perception Engineer

HorseShoe Beverage Company; South San Francisco (CA); Hybrid; USD 150,000 - 200,000; Full time

Job summary

BurnBot is seeking an experienced engineer to own the perception stack for remotely operated machines used in wildfire management. You will integrate camera, lidar, thermal, and radar data, calibrate sensors, and fuse outputs to detect vegetation, terrain, obstacles, and people in real environments.

You will deploy perception models on embedded hardware, optimize for latency and power, and evaluate performance with real-field metrics, contributing to safe, rapid wildfire response.

Qualifications

• 5+ years building perception for robots or vehicles outside a lab.
• Strong C++ and Python programming.
• Deep learning for detection and segmentation.
• Multi-sensor calibration and fusion expertise.
• ROS or ROS 2 experience and field data tooling.

Responsibilities

• Own the perception stack: sensors, models, and data engine for field robots.
• Develop detection/segmentation for vegetation, terrain, obstacles, fire, and people.
• Ensure deployment onto embedded compute with real latency and power constraints.
• Provide honest evaluation with real metrics on field data.

Skills

C++

Python

Deep learning

Multi-sensor fusion

ROS/ROS 2

Field data tooling

Education

B.S. or higher in robotics, CS, or EE

Tools

Bags

Replay

Visualization

Job description

• Location 340 Shaw Rd,South San Francisco, CA, 94080,United States
• Base Pay $150,000.00 - $200,000.00 / Year
• Employee Type Exempt (FT)

Description

About BurnBot

Wildfire has stopped being a season. BurnBot builds the machines that get ahead of it — remotely operated systems that apply and contain prescribed fire, and platforms that clear heavy fuel on terrain that isn’t safe for crews.

Our fleet is at work today for PG&E, CAL FIRE, the U.S. Forest Service, Caltrans, fire districts, and tribal nations across California, Oregon, and Washington — treating land at up to 40x the pace of conventional methods. We’re roboticists, engineers, and career fire practitioners. Hardware we design in South San Francisco is in the dirt within weeks, and the people who designed it are usually standing next to it.

THE ROLE

Our machines have to see what a fire crew sees: where the fuel is, how heavy it is, where the ground drops away, where the fire line is, and where the people are. They have to do it through smoke, dust, glare, and at night.

You’ll own the perception stack that makes that possible — sensors, models, and the data engine behind them — on machines that work in real vegetation, not a parking lot.

WHAT YOU’LL OWN

The perception stack: camera, lidar, thermal, and radar — selection, placement, calibration, and fusion

Detection and segmentation of vegetation, terrain, obstacles, fire, and people, tuned for the conditions our machines actually work in

The data engine: what we collect in the field, how it gets labeled, and how quickly a failure becomes a training example

Evaluation you trust. Real metrics on real field data, and the honesty to say when a model isn’t ready.

Deployment onto embedded compute on the machine, within real latency and power budgets

Requirements

WHAT YOU’LL BRING

5+ years building perception for robots or vehicles that operate outside a lab

Strong C++ and Python

Deep learning for detection and segmentation, and the judgment to know when classical methods are the better answer

Multi-sensor calibration and fusion, and comfort with the geometry underneath it

ROS or ROS 2, and the tooling around field data — bags, replay, visualization

Field debugging instincts. You’ve chased a bug that only reproduced outdoors

B.S. or higher in robotics, computer science, electrical engineering, or equivalent experience

NICE TO HAVE

Thermal or IR imaging experience

Lidar in vegetation — foliage penetration, ground segmentation, or point-cloud classification

Off-road or unstructured-environment perception (agriculture, mining, forestry, defense)

Active learning or data-engine tooling at scale

Wildland fire, forestry, or land management background

Summary

HOW WE WORK

Based onsite in South San Francisco — design, build, and test all happen in one building, and being near the machines is most of the point

Flexible within that. You own your calendar: work the hours where you’re sharpest, and go remote on heads-down days.

Expect a few days a month in the field across California, Oregon, and Washington. Watching your work run on a live burn is the best part of the job.

COMPENSATION & BENEFITS

Base salary: $150,000 – $200,000 — final offers reflect experience and demonstrated scope.

Burnbot is an Equal Opportunity Employer and does not discriminate based upon race, religion, color, national origin, sex (including but not limited to, pregnancy, childbirth, or related medical conditions), sexual orientation, gender, gender identity, gender expression, transgender status, sexual stereotypes, age, status as a protected veteran, status as an individual with one or more disabilities, or other applicable legally protected characteristics.

If you need assistance or accommodation due to a disability, you may contact us at [phone] x 1010 or [e-mail].