Liquid AI is a Cambridge-based AI lab building general-purpose models engineered for efficiency. Its stack pairs training frameworks (PyTorch, JAX, DeepSpeed, Megatron-LM) with low-level GPU tooling (CUDA, cuBLAS, Nsight) and edge runtimes like llama.cpp and ONNX Runtime, pointing to deep investment in inference performance under tight latency and memory limits. Hiring is accelerating and skews heavily toward senior engineering talent.
Liquid AI develops general-purpose AI models designed to run efficiently across a range of deployment sizes. Founded in 2023 and headquartered in Cambridge, Massachusetts, the privately held company employs 51–200 people and hires in the United States. Its work spans model training, custom GPU kernel development, inference optimization and audio model adaptation for enterprise customers. A web-facing product layer is built with React, Next.js, TypeScript and Supabase, alongside a growing solutions architecture function supporting customer deployments.
Liquid AI uses PyTorch, JAX, DeepSpeed, Megatron-LM, CUDA, cuBLAS, NCCL, vLLM, SGLang, TensorRT, llama.cpp and ONNX Runtime, plus C++, Rust and Python, with React, Next.js and Supabase on the web side.
Yes. Hiring at Liquid AI is accelerating and concentrated in engineering, mostly at senior level, with additional openings in research, data, product, legal and support across the United States.
Current work includes an end-to-end model development pipeline, benchmark harnesses for performance regression, custom CUDA kernels for novel model architectures, and data generation pipelines for audio training.
Other companies in the same industry, closest in size
Liquid AI's technology stack, projects, and hiring signals are inferred from public hiring and company data — career pages, public listings, and company web presence — then clustered and de-duplicated. Figures are estimates that refresh over time. Read our full methodology →
This is not an official vendor or customer list. It is a technology-adoption signal inferred from public data, intended for B2B research.
