About The Position

This role involves designing, training, and deploying real-time vision models for detection, classification, tracking, and state estimation on fast-moving aerial targets. The engineer will optimize networks and CV algorithms using CUDA and TensorRT for real-time performance on NVIDIA Jetson Orin-class hardware. Integration into the autonomy stack using ROS 2 and writing production-quality C++ for real-time processing are key aspects. The position also includes deploying, benchmarking, and maintaining models on embedded edge compute aboard autonomous vehicles, as well as participating in field tests and data collection to rapidly iterate on model development.

Requirements

  • B.S. in Computer Science, Robotics, Electrical Engineering, or a related field plus 1+ years of industry or applied robotics experience, or an M.S. in CS, Robotics, Computer Vision, or a related discipline.
  • Experience training and fine-tuning vision models (CNNs, Vision Transformers, detection/tracking).
  • Proficiency in C++ for performance-sensitive systems.
  • Hands-on GPU programming or profiling experience (CUDA preferred; coursework or projects count).

Nice To Haves

  • Deploying models on embedded edge platforms (Jetson Orin/Xavier).
  • ROS 2 experience (nodes, lifecycle management, QoS, custom interfaces).
  • Sensor fusion (visual-inertial odometry, camera + radar/LiDAR).
  • Prior field-robotics or UAS test experience.

Responsibilities

  • Design, train, and deploy real-time vision models for detection, classification, tracking, and state estimation on fast-moving aerial targets.
  • Optimize networks and CV algorithms with CUDA and TensorRT to hit end-to-end latency on NVIDIA Jetson Orin-class hardware.
  • Integrate perception into our autonomy stack using ROS 2 and write production-quality C++ for real-time processing.
  • Deploy, benchmark, and maintain models on embedded edge compute aboard autonomous vehicles.
  • Join field tests and data collection, then turn what you see on the range into the next model iteration within days, not quarters.

Benefits

  • Real hardware, real flights, fast iteration loops. Your code leaves the laptop.
  • Small team, direct ownership, and visibility with leadership.
  • Meaningful work on a national-security problem.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service