Real-Time Perception Engineer for Robotic Vision & Control

Path Robotics; Columbus (OH); On-site; USD 130,000 - 170,000; Full time

Benefits offered by this job

Free lunch

Flexible PTO

Medical, dental, vision

Parental leave

401(k) retirement plan

Employee referral bonuses

Job summary

Path Robotics is seeking an exceptional hands-on Perception Engineer to own the vision and real-time control loop on embedded hardware, delivering robust perception from sensing to output acted on by another engineer's control loop.

You will work on the end-to-end perception pipeline, handle degraded imagery, and validate estimates with proper uncertainty. Requirements include modern production C++ and ROS 2, plus GPU-accelerated inference on embedded hardware within a fixed per-frame budget.

Qualifications

• Real-time loop on camera data with physical movement and measurable latency.
• First-principles depth in classical vision and 3-D geometry including calibration and error propagation.
• Real-time segmentation or feature extraction on degraded imagery, deployed in production.
• Strong production C++ and ROS 2 experience.
• GPU-accelerated inference on embedded hardware with per-frame budget considerations.

Responsibilities

• Develop end-to-end real-time perception pipeline from sensing to embedded control.
• Ensure robust measurements when signals are degraded.
• Make estimates with explicit failure handling and uncertainty.
• Perform geometric estimation against known models with partial data.
• Evaluate sensing options and defend choices with data.
• Help shape technical direction toward learned methods.

Skills

C++

ROS 2

Real-time systems

3-D geometry

Camera geometry

Tools

ROS 2

Job description

Path Robotics is seeking an exceptional hands-on Perception Engineer to own the vision and real-time control loop on embedded hardware, delivering robust perception from sensing to output acted on by another engineer's control loop.

You will work on the end-to-end perception pipeline, handle degraded imagery, and validate estimates with proper uncertainty. Requirements include modern production C++ and ROS 2, plus GPU-accelerated inference on embedded hardware within a fixed per-frame budget.