Machine Learning Engineer at Liquid

Setup: Remote, Hybrid
Location: San Francisco, California
Type: Full-time
Posted: Dec 16, 2025

Description

Liquid, spun out of MIT, is redefining AI architecture by building efficient AI systems at every scale, particularly Liquid Foundation Models (LFMs) for on-device, edge, and real-time applications. The team is comprised of world-class engineers and researchers focused on shaping model architectures, scaling dev platforms, and enabling enterprise deployments. The role involves inventing new model architectures for inference speed, building evaluation suites, collaborating on scalable data pipelines, optimizing model training, contributing to research, and working with applied research and business teams on client use cases.

Required Skills

Machine learning
Audio models
PyTorch
Distributed training
Multimodal data
Data quality
Model architectures
Evaluation suites
Python

Additional Information

Required Previous Experiences: Experience with machine learning at scale
Worked with audio models and understand the effects of architecture choices on runtime, latency, and quality
Proficient in PyTorch, and familiar with distributed training frameworks like DeepSpeed, FSDP, or Megatron-LM
Worked with multimodal data (e.g. audio, text, image, video)
Contributed to research papers, open-source projects, or production-grade multimodal model systems
Understand how data quality, augmentations, and preprocessing pipelines can significantly impact model performance—and built tooling to support that
Enjoy working in interdisciplinary teams across research, systems, and infrastructure, and can translate ideas into high-impact implementations
Preferred Qualifications: Designed and trained multimodal language models or specialized audio models (e.g., ASR, TTS, voice conversion, vocoders, diarization)
Deep care about empirical performance, and ability to design, run, and debug large-scale training experiments on distributed GPU clusters
Developed audio encoders or decoders, or integrated them into language pretraining pipelines with autoregressive or generative objectives
Experience working with large-scale audio datasets, understanding unique challenges, and effective management
Strong programming skills in Python, with emphasis on clean, maintainable, and scalable code