Member of Technical Staff - Research Engineer, Omni

Skills

About the Role

You will run end-to-end multimodal model training for Omni, covering data, pre- and post-training, evals, inference and deployment. You will contribute to open research such as multimodal MoE and token-efficient encoders, help consolidate training infrastructure, and turn research-grade code into maintainable, production-ready systems while working across vision and audio teams.

Requirements

• Hands-on experience training ML models in at least one non-text modality (vision most likely; audio can be learned)
• Foundational understanding of how multimodal training works below the research layer
• Comfortable between research and engineering, running experiments and hardening code
• Able to take training infrastructure and inference from experiment code to production-ready
• Preferred: audio and speech modeling, inference/deployment (including on-device), multimodal MoE exposure, publications

Responsibilities

• Run multimodal training experiments end to end, pre- and post-training, and fold results into the next Omni release
• Contribute to open research directions such as multimodal MoE and token efficiency for multimodal encoders
• Share multimodal data work: pipelines, curation, and data decisions
• Harden research-grade experiment code into maintainable infrastructure
• Help consolidate Omni's training infrastructure into one codebase and library
• Own inference and deployment for Omni models
• Work across vision and audio teams on data, infra, training, and evals

Benefits

• Equity
• Unlimited PTO
• Company-wide Refill Days
• Comprehensive health coverage
• Paid leave for pregnancy, parenting, medical needs, and family care