Machine Learning Engineer — Multimodal & Vision

Role overview Build and deploy machine learning systems that speed up the preparation of vision datasets for autonomous driving and robotics. The work spans model fine-tuning, efficient GPU inference, confidence evaluation against human labels, and integration with customer machine learning operations systems.

Responsibilities - Fine-tune and deploy assisted labeling models, including segmentation, object detection, and 3D point cloud approaches. - Improve GPU inference speed using tools such as TensorRT, ONNX, and Triton. - Compare model confidence with human consensus to improve active learning and review workflows. - Integrate model-assisted annotation APIs into customer MLOps infrastructure.

Requirements - At least three years building and deploying production deep learning models. - Experience with PyTorch, Docker, and CUDA. - Expertise in computer vision, 3D LiDAR or sensor fusion, or multimodal vision-language systems. - Strong Python and cloud engineering skills, including AWS or GCP. - Experience optimizing inference for real-time interactive applications.

Benefits and work setup - Remote role based in the United States, with flexible hours across time zones. - Salary range of $140,000–$185,000 plus equity. - Cloud GPU budget for experimentation, healthcare coverage with mental wellness support, and a hardware allowance.