Staff ML Research Engineer
Core
Define architectures and training strategies for next-generation multimodal embedding models, building large-scale data curation systems and production APIs for video, audio, text, and document assets.
Role type
Staff ML Research Engineer (multimodal)
Builds
Production multimodal embedding models, distributed training pipelines, and large-scale data curation systems
Domain
Computer Vision, Video Understanding, Multimodal Learning
Deliverable
production ML models
Required skills
Large-scale distributed model training, Video understanding, Multimodal representation learning, Production model API development, Python, PyTorch, Model architecture design, Data curation and quality systems, GPU utilization optimization
Preferred skills
Billion-parameter scale training, Temporal/spatial/hierarchical modeling, Multimodal video modeling, Training and inference optimization, Technical leadership
Technologies
PyTorch, FSDP, GPU clusters
Responsibilities
Set technical direction for multimodal structure and define reusable addressable units for video, audio, text, and documents; Architect and optimize distributed training pipelines and data-processing systems; Own production model APIs including packaging, API design, and inference optimization; Build large-scale data curation, filtering, and quality systems; Define evaluation methods and quality standards for multimodal understanding and retrieval; Drive interface between multimodal structure and semantics aligning with Agent, Search, Product, and Infrastructure teams; Raise the research engineering bar through design reviews and technical mentorship.
Seniority
Staff, hands-on IC with strategic direction