Senior, ML Engineer - VLM
Core
Design, implement, and deploy cloud-based pipelines to convert logged multi-sensor fleet data into VLM/VLA training datasets, including geometric, semantic, and reasoning-grounded annotations.
Role type
Senior IC machine-learning engineer (VLM data curation)
Builds
VLM/VLA-ready datasets (geometric labels, semantic descriptions, reasoning traces) for end-to-end autonomous driving models
Domain
Autonomous driving / Computer Vision / Multimodal AI
Deliverable
production ML models
Required skills
Computer Vision (2D/3D detection, tracking, BEV, depth estimation), Deep Learning, Vision-Language Models (open-vocabulary, dense captioning), Large-scale data processing (Parquet), Distributed ML frameworks (PyTorch, Ray, Spark), MLOps (MLflow, W&B), Python, Cloud infrastructure (AWS)
Preferred skills
VLA models, Auto-labeling foundation models, High-throughput model serving (vLLM), Semantic retrieval (vector databases), AV data standards (ROS, MCAP, Pegasus), Data visualization (Foxglove, FiftyOne), Research publications
Responsibilities
Design and deploy cloud pipelines for multi-sensor data conversion, develop VLM-assisted auto-labeling systems, generate reasoning-grounded labels, mine and curate long-tail failure cases, define dataset schemas and quality metrics, partner with model teams on specifications, scale distributed pipelines, lead and mentor engineers
Seniority
Senior, hands-on IC with mentorship responsibilities
