Member of Technical Staff Multi Modal Vision
Liquid AI is an efficiency-first foundation-model company building device-native Liquid Foundation Models (LFMs) and tools to customize and deploy them.
Funding history
About Liquid AI
An MIT CSAIL spinout, Liquid AI develops general-purpose AI models focused on efficient deployment across CPUs, GPUs, NPUs, edge devices, and cloud or on-premises environments.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will lead model capabilities from task specification through data curation, training, ablations, evaluation, and deployment. You will research and implement scalable vision-language model improvements, improve visual reasoning with reinforcement learning and preference optimization, and optimize token efficiency through encoder and connector design.
Requirements
- Experience training or evaluating vision-language models
- Experimental rigor
- Ability to turn research ideas into scalable implementations
- Python proficiency
- Proficiency with at least one deep learning framework
- Experience refining and iterating through hypotheses
Responsibilities
- Lead model capabilities from task specification through deployment
- Develop data curation, training recipes, ablations, and evaluations
- Improve visual reasoning through reinforcement learning and preference optimization
- Optimize token efficiency through encoder and connector design
- Contribute to production vision-language models
Benefits
- Equity
- Medical, dental, and vision premiums fully paid for employees and dependents
- 401(k) matching up to 4% of base pay
- Unlimited PTO
- Company-wide Refill Days