ICML 2026

Liquid AI • ,

3d

About The Position

Liquid AI, spun out of MIT CSAIL, is building foundation models from scratch using a fundamentally different hybrid architecture called Liquid Foundation Models (LFMs). These models offer faster inference, lower memory usage, and can be deployed in environments where traditional models cannot. The company releases open-weight text, vision-language, and audio-language models designed to run on various devices, including phones, laptops, vehicles, and embedded systems.

Requirements

• Demonstrated research or engineering contribution in one or more of the specified areas (efficient architectures, multimodal vision, multimodal audio, data engineering, infrastructure & performance, post-training & alignment).
• Ability to move from idea to implementation to shipped result.
• M.S. or Ph.D. in Computer Science, Mathematics, Electrical Engineering, or a related field; or equivalent industry experience.

Nice To Haves

• Published research at top-tier venues (NeurIPS, ICML, ICLR, CVPR, ACL, Interspeech, etc.).
• Experience training or fine-tuning foundation models at scale.
• Hands-on work with distributed training infrastructure (DeepSpeed, FSDP, Megatron-LM).
• Open-source contributions (code, data, or models) on GitHub or HuggingFace.
• Experience deploying models to edge or on-device environments.

Responsibilities

• Contribute to research and engineering in areas such as efficient architectures, multimodal vision and audio, data engineering, infrastructure & performance, and post-training & alignment.
• Develop and implement novel model architectures, including state space models, hybrid attention designs, neural ODEs, and alternatives to the transformer paradigm.
• Work on vision-language models that operate on-device with strict latency and memory constraints.
• Develop speech foundation models, end-to-end audio-language systems, and real-time voice solutions for constrained hardware.
• Curate pre-training data, generate synthetic data, and develop data mixture and scaling strategies.
• Optimize distributed training, GPU kernels, edge inference, and model serving at scale.
• Implement and refine post-training techniques such as RLHF, preference optimization, and multi-stage reinforcement learning.
• Own workstreams end-to-end, from research and ideation through implementation and shipped models.
• Publish research findings and release open-weight models.
• Present work at conferences.

Benefits

• Full ownership of work from architecture to deployment.
• Competitive base salary with equity.
• 100% coverage of medical, dental, and vision premiums for employees and dependents.
• 401(k) matching up to 4% of base pay.
• Unlimited PTO.
• Company-wide Refill Days.
• Visa sponsorship (O-1 and H-1B) for exceptional talent.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior