Senior Computer Vision Engineer — On-Edge Targeting

Build real-time detection and tracking for thermal (LWIR) and electro-optical pipelines that hold small aerial and ground targets through motion, occlusion, and smoke.

Targeting in the real world means holding a small, fast, partly hidden object through motion blur, smoke, and thermal clutter — at frame rates that keep fire-control honest. You will build and harden the detection-and-tracking stack across LWIR and electro-optical sensors, then make it run inside tight Jetson power and latency budgets. Your models are validated against captured field video and tuned until frame drop stops costing engagements. This is applied, measurable work where a few milliseconds and a few percentage points genuinely matter.

• Strong background in detection, optical-flow tracking, and sensor fusion
• Expert in TensorRT, CUDA C++, OpenCV, and GStreamer
• Familiarity with ByteTrack/CSRT tracking and YOLO/RT-DETR detection
• Rigorous testing against realistic field video

• Maintain the targeting tracker inside low-latency Jetson runtimes
• Benchmark inference against high-framerate feeds
• Implement multi-spectral (thermal + EO) fusion
• Minimize frame drop for reliable fire-control