Foundation Model Engineer

Join Liquid AI to pioneer the next generation of generative AI using Liquid Neural Networks (LNNs) and bio-inspired dynamical systems. As a Foundation Model Engineer, you will design and scale efficient, state-of-the-art models that surpass the limitations of traditional Transformers in memory footprint and inference speed. Your work will directly contribute to building general-purpose, high-performance foundation models capable of handling massive context windows and real-time reasoning.

- Design and implement novel neural architectures based on Liquid Neural Networks and continuous-time dynamical systems. - Scale foundation models to billions of parameters while maintaining computational efficiency and stability. - Develop and optimize high-throughput training pipelines for multi-node GPU clusters. - Conduct rigorous empirical evaluations and ablation studies to benchmark LNNs against industry-standard Transformers and SSMs. - Collaborate with the systems team to integrate hardware-aware optimizations for edge and cloud deployment. - Translate cutting-edge AI research into production-ready code that powers our core API offerings. - Participate in peer code reviews and contribute to the company's long-term technical roadmap for efficient AI.

- Master’s or PhD in Computer Science, Physics, or Mathematics with a focus on Deep Learning. - Proven track record of training large-scale foundation models (LLMs, SSMs, or RNNs) from scratch. - Deep theoretical understanding of dynamical systems, differential equations, or sequence modeling architectures. - Expert-level proficiency in PyTorch and experience with distributed training frameworks like DeepSpeed or Megatron-LM. - Experience writing custom CUDA kernels or Triton code to optimize non-standard architectural layers. - History of contributing to high-impact research papers or major open-source machine learning projects. - Ability to work independently in a fully remote, globally distributed team across multiple time zones.