4 hours ago
Boston, MA, USAMid Level / Senior
H1B sponsor
Responsibilities
- Run multimodal training experiments end to end, including pre-training, post-training, evaluation, and release integration.
- Contribute to research directions such as multimodal mixture-of-experts and token-efficient multimodal encoders.
- Build and manage multimodal data pipelines, curation, and release-related data decisions.
- Harden research-grade experiment code into maintainable team infrastructure.
- Consolidate Omni’s training infrastructure, including distributed-training choices and a unified codebase and library.
- Own inference and deployment for Omni models and collaborate across vision and audio teams.
Requirements
- Hands-on experience training ML models in at least one non-text modality, especially vision or audio.
- Foundational understanding of multimodal training based on prior hands-on work.
- Ability to operate between research and engineering by running experiments and hardening supporting code.
- Experience taking training infrastructure and inference from experimental code to maintained, production-ready systems.
- Strong interest or experience in training infrastructure, inference, or rebuilding and consolidating training stacks.
- Audio or speech modeling experience, on-device inference, multimodal mixture-of-experts or encoder token-efficiency research, and publications are helpful but not required.
Benefits
- Competitive base salary plus equity.
- Unlimited paid time off and company-wide Refill Days.
- Comprehensive health coverage, with details varying by location.
- Paid leave for pregnancy, parenting, medical needs, and family caregiving.
- Full-time hybrid work in Boston.
Categories
AI ResearchML Engineering
About Liquid AI
Liquid AI builds general-purpose AI systems that run efficiently from data center accelerators to on-device hardware, emphasizing low latency, memory efficiency, privacy, and reliability. The company partners with enterprises in consumer electronics, automotive, life sciences, and financial services to deploy and benchmark models for real-world workloads. Founded in 2023 out of MIT CSAIL and headquartered in Cambridge, Massachusetts, it is privately held.