About the job
About Liquid AI
Liquid AI, a pioneering company spun out of MIT CSAIL, is at the forefront of developing general-purpose AI systems that operate efficiently across various platforms, from data center accelerators to on-device hardware. Our commitment to low latency, minimal memory usage, privacy, and reliability allows us to partner with some of the most esteemed enterprises in consumer electronics, automotive, life sciences, and financial services. As we experience rapid growth, we are seeking exceptional talent to join our innovative journey.
The Opportunity
Join our cutting-edge Audio team, where we are developing advanced speech-language models capable of handling Speech-to-Text (STT), Text-to-Speech (TTS), and speech-to-speech tasks within a unified architecture. This pivotal role supports applied audio model development, directly collaborating with the technical lead to deliver production systems that operate on-device under real-time constraints. You will take ownership of key workstreams encompassing data pipelines, evaluation systems, and customer deployments. If you are eager to tackle unique technical challenges within a small, elite team where your contributions are impactful, this is the role for you.
What We're Looking For
We are seeking an individual who:
- Builds first, theorizes later: You prioritize shipping working systems over theoretical models; production-grade code is your default.
- Owns outcomes end-to-end: You take full responsibility for everything from data pipelines to customer deployments and don't shy away from challenges.
- Thrives under constraints: On-device, low-latency, memory-constrained environments motivate you. You view constraints as opportunities for innovative design.
- Ramps quickly on new territory: You are comfortable closing knowledge gaps swiftly and actively seek feedback to drive results.
The Work
- Develop and scale data pipelines for audio model training, including preprocessing, augmentation, and quality filtering at scale.
- Design, implement, and maintain evaluation systems that assess multimodal performance across both internal and public benchmarks.
- Fine-tune and adapt audio models to cater to customer-specific use cases, taking charge from requirement gathering through to deployment.
- Contribute production code to the core audio repository while collaborating closely with infrastructure and research teams.
- Facilitate experimentation under real hardware constraints, transitioning smoothly between customer-focused projects and core development initiatives.
