Evan Brossard

About Evan Brossard

I build perception systems that help machines see and understand the world.As a Senior Deep Learning Engineer at Carbon Robotics, I develop real-time computer vision pipelines for autonomous agricultural tractors. My work involves designing and deploying semantic segmentation models, object detection systems, and multi-sensor fusion algorithms that run on NVIDIA edge platforms.What drives me is the challenge of taking cutting-edge research and making it work in the real world—where milliseconds matter, hardware is constrained, and conditions are unpredictable.My experience spans:• Autonomous systems and robotics perception• Real-time model optimization (PyTorch → TensorRT → CUDA)• Multi-modal sensor fusion (LiDAR, radar, cameras)• Edge AI deployment on embedded platformsBefore autonomous tractors, I worked on volumetric telepresence systems at Omnivor, where I optimized 3D reconstruction pipelines and real-time video streaming.

Experience

1. Senior Deep Learning Engineer

   Carbon Robotics

   Mar 2021 — Present · Seattle, WA, US

   Architect and owner of real-time perception pipeline integrating multi-modal sensor fusion (LiDAR, radar, cameras) with high-throughput inference on NVIDIA edge GPUs- Designed and implemented novel furrow detection model from research through TensorRT deployment, enabling autonomous row following with sub-centimeter accuracy- Led migration to TensorRT 8.x across production fleet, optimizing inference performance and maintaining model accuracy while reducing latency by 10%- Engineered high-throughput post -processing pipeline supporting detection points per frame at 10 FPS using optimized CUDA kernels and efficient data structures- Drove camera system reliability improvements, collaborating with hardware vendors to resolve firmware issues and implementing software-based connection fault tolerance- Mentored engineers and interns on deep learning best practices, model optimization, and real-time deployment strategies for edge AI applications- Inventor on 2+ patents including autonomous turnaround routines and weed/crop detectionmethodologies

Education

• University of Washington

  Bachelor of Arts (B.A.), Computer Science

Skills

• C#
• Game Design
• Html5
• 3d Studio Max
• Unity3d
• Video Games
• Visual Studio
• Linux
• Mysql
• Powerpoint
• Javascript
• Html 5
• Game Development
• Scripting
• Java
• Microsoft Office
• Programming