11 months ago
Responsibilities
- Design and implement scalable ML training pipelines for computer vision models covering object detection, tracking, classification, and segmentation.
- Build efficient model-serving infrastructure for real-time inference on edge devices with constrained compute and power budgets.
- Optimize models for embedded deployment using quantization, pruning, TensorRT, ONNX, and CoreML.
- Develop continuous training and evaluation systems using production data feedback loops.
- Create pipelines for ingesting, labeling, versioning, and managing video, radar, lidar, and thermal sensor datasets.
- Implement model monitoring, A/B testing frameworks, and performance analytics for deployed perception systems.
- Collaborate with perception researchers to transition models from research to production across thousands of edge nodes.
- Build infrastructure and tools for distributed training, hyperparameter optimization, and experiment tracking.
Requirements
- Strong experience with PyTorch and TensorFlow and model optimization tools including TensorRT, ONNX Runtime, and OpenVINO.
- Deep understanding of computer vision architectures and deployment tradeoffs, including YOLO, transformers, CNNs, and real-time detection and tracking.
- Hands-on experience deploying models on edge devices such as NVIDIA Jetson and ARM processors or similar embedded platforms.
- Expertise building MLOps infrastructure with experiment tracking, feature stores, model registries, and CI/CD for ML.
- Experience with distributed training frameworks including PyTorch DDP, DeepSpeed, and Ray, as well as GPU cluster management.
- Strong software engineering skills in Python and systems languages such as C++ and Rust for performance-critical inference code.
- Familiarity with video processing, sensor fusion, or multimodal perception systems is preferred.
- Prior experience in robotics, autonomous systems, or real-time ML applications is highly valued.
About Specter
Specter delivers real-time data and insights on private companies, enabling investors to make informed, confident decisions. Harness the power of live data and AI-driven analysis to outsmart the competition and make confident decisions in private markets.