Job Description

• Take pre-trained vision models and optimize them for deployment on NVIDIA embedded devices using TensorRT and related tooling
• Build and maintain the perception pipeline: depth cameras, point cloud processing, object detection using off-the-shelf sensor models
• Evaluate perception methods. Run experiments, measure performance, make pragmatic decisions about what to use
• Integrate OpenCV and PyTorch-based inference into the broader ROS 2 system
• Work with ZED stereo cameras and similar sensors; exploit bundled SDK models where they're good enough

• Strong Python; comfortable with PyTorch and OpenCV in a production/deployment context
• Hands-on experience with TensorRT or other model optimization/deployment tools (ONNX, DeepStream, etc.)
• Perception fundamentals: understanding of depth sensing, point clouds, basic object detection
• Comfortable evaluating and benchmarking approaches, not just implementing specified ones
• Familiarity with visual SLAM or similar sensor fusion techniques is a strong plus

For applications and inquiries, contact: hirings@openkyber.com

Job Tags

Similar Jobs