Member of Technical Staff, Multimodal Vision
Core
Building multimodal foundation models (text, video, vision, audio) and training infrastructure for agentic AI systems.
Role type
Senior IC multimodal machine learning engineer (research & infrastructure)
Builds
Multimodal foundation models, RL training stacks, evaluation frameworks, and distributed training systems
Domain
Artificial Intelligence, Multimodal Learning, Reinforcement Learning
Deliverable
production ML models
Required skills
Multimodal model development, Mid-training and post-training (SFT, RLHF/RLAIF, RLVR), Reward modeling, Preference optimization, Distributed training systems, Long-context video understanding, Data-driven experimentation
Preferred skills
RL training infrastructure scaling, Video-audio or speech-language models, Streaming real-time multimodal systems, Top-tier publications or open-source contributions
Technologies
Megatron-LM, FSDP, DeepSpeed, PPO, GRPO, DPO
Responsibilities
Drive mid-training and post-training research for multimodal foundation models; Own the RL stack for multimodal post-training; Build multimodal evaluation frameworks and internal benchmarks; Build and optimize multimodal training infrastructure
Seniority
Senior, hands-on IC