Member of Technical Staff - Machine Learning Engineer, Inference (Pytorch)

Liquid AI, an MIT spin-off, is a foundation model company headquartered in Boston, Massachusetts. Our mission is to build capable and efficient general-purpose AI systems at every scale.

Our goal at Liquid is to build the most capable AI systems to solve problems at every scale, such that users can build, access, and control their AI solutions. This is to ensure that AI will get meaningfully, reliably and efficiently integrated at all enterprises. Long term, Liquid will create and deploy frontier-AI-powered solutions that are available to everyone.

We are hiring an ML Engineer (Inference) to build and optimize the end-to-end serving stack for Liquid AI’s foundation models. You will develop the pipeline between a trained model checkpoint and a production-grade, low-latency API. This is a highly technical role operating on the frontier of AI inference research and production

• PyTorch
• Python
• Model-serving frameworks (e.g. TensorRT, vLLM, SGLang)

• You have experience building large-scale production stacks for model serving.
• You have a solid understanding of ragged batching, dynamic load balancing, KV-cache management, and other multi-tenant serving techniques.
• You have experience with applying quantization strategies (e.g., FP8, INT4) while safeguarding model accuracy.
• You have deployed models in both single-GPU and multi-GPU environments and can diagnose performance issues across the stack.

• Optimize and productionize the end-to-end pipeline for GPU model inference around Liquid Foundation Models (LFMs).
• Facilitate the development of next-generation Liquid Foundation Models from the lens of GPU inference.
• Profile and robustify the stack for different batching and serving requirements.
• Build and scale pipelines for test-time compute.

• Hands-on experience with state-of-the-art technology at a leading AI company.
• Deeper expertise in machine learning systems and efficient large model inference.
• Opportunity to scale pipelines that directly influence user latency and experience with Liquid's models.
• A collaborative, fast-paced environment where your work directly shapes our products and the next generation of LFMs.