Liquid AI

Liquid AI, a rapidly scaling company spun out of MIT CSAIL, is building efficient, general-purpose AI systems for diverse enterprise applications. This role offers a unique opportunity to work at the forefront of vision-language models (VLMs) and their real-world deployment. You will take end-to-end ownership of applied VLM post-training for major enterprises, while also contributing to the company's core multimodal model development. This position allows you to shape the future of applied multimodal AI by focusing on visual understanding, data quality, evaluation, and production readiness.

Liquid AI, a rapidly scaling company spun out of MIT CSAIL, is seeking an Applied ML professional to join their team. This role offers a unique opportunity to apply advanced sequential recommendation architectures to real-world enterprise problems at scale. You will be responsible for the end-to-end applied ML work for recommendation system workloads, adapting large-scale models for customers requiring efficient personalization and ranking capabilities under production constraints. This position provides full ownership over model adaptation, evaluation, and deployment, differing from more siloed recommendation roles. You will also contribute to building reusable tooling and workflows to accelerate future projects, making this an ideal role for those passionate about data quality, user behavior modeling, and implementing effective enterprise recommendation systems.

Liquid AI, a company spun out of MIT CSAIL, is developing general-purpose AI systems designed for efficient deployment across various targets, from data centers to on-device hardware, prioritizing low latency, minimal memory usage, privacy, and reliability. They are expanding rapidly and seeking exceptional talent to join their team. This role focuses on "LFM2.5-Audio," Liquid's multimodal speech and text language model, which handles speech-to-speech conversation, ASR, and TTS without separate components, making it ideal for real-time, on-device applications. The core challenge for this position is adapting audio models to understand user intents and translate them into structured tool calls, enabling voice-driven function calling for enterprise customers. This position offers a unique opportunity to work at the intersection of advanced audio models and real-world deployment, owning the applied post-training work from data generation to delivery, and directly influencing the evolution of Liquid's post-training and audio stacks.

Liquid AI, spun out of MIT CSAIL, builds general-purpose AI systems that run efficiently across various deployment targets. We partner with enterprises in consumer electronics, automotive, life sciences, and financial services. This role offers a unique opportunity to lead applied post-training work end-to-end for text workloads, adapting Liquid Foundation Models for major enterprise clients. You will serve as the technical liaison between customer needs and model delivery, owning engagements from scoping to evaluation. Between projects, you will develop reusable applied workflows and tooling to enhance future delivery. This position is ideal for someone passionate about data quality, evaluation design, and making language models effective in production environments.

Liquid AI, a rapidly scaling company spun out of MIT CSAIL, is building general-purpose AI systems designed for efficiency across various deployment targets, from data centers to on-device hardware. These systems prioritize low latency, minimal memory usage, privacy, and reliability for enterprise clients in sectors like consumer electronics, automotive, life sciences, and financial services. The Audio team is developing advanced speech-language models for STT, TTS, and speech-to-speech applications within a unified architecture. This role focuses on applied audio model development, working closely with the technical lead to deliver production systems that operate under real-time, on-device constraints. The position offers high ownership of critical workstreams, including data pipelines, evaluation systems, and customer deployments, within a small, elite team where code directly ships to production.

Liquid AI, a rapidly scaling company spun out of MIT CSAIL, is developing general-purpose AI systems designed for efficiency across various deployment targets. The Data team is crucial for powering Liquid's Foundation Models, focusing on creating high-quality, purpose-built datasets as public data sources become insufficient. This role involves treating data as a research problem, where engineers design experiments, run ablations, and measure the impact of data decisions on model quality. You will be matched to a team focusing on pre-training, post-training RL, vision-language, audio, or multimodal applications.

Liquid AI, spun out of MIT CSAIL, is developing general-purpose AI systems designed for efficient execution across various deployment targets, from data center accelerators to on-device hardware. The company focuses on ensuring low latency, minimal memory usage, privacy, and reliability, partnering with enterprises in consumer electronics, automotive, life sciences, and financial services. The Training Infrastructure team is building the distributed systems that power next-generation Liquid Foundation Models, requiring the design, implementation, and optimization of large-scale training infrastructure. This role is focused on runtime, performance, and reliability within training systems, offering a high-ownership opportunity to build critical systems from the ground up within a small team with fast feedback loops.

Liquid AI, a rapidly scaling company spun out of MIT CSAIL, is developing general-purpose AI systems designed for efficient deployment across various hardware, from data centers to edge devices. We focus on delivering low latency, minimal memory usage, privacy, and reliability for enterprise clients in sectors like consumer electronics, automotive, life sciences, and financial services. We are seeking exceptional individuals to join our small, fast-moving, high-ownership team to tackle complex performance challenges that generic frameworks cannot address.