Together AI

Together AI builds a purpose-built GPU cloud platform that lets developers and organizations of any size train, fine-tune, and deploy generative AI models without vendor lock-in. The company sits at the intersection of AI infrastructure and generative AI, working across distributed systems, distributed inference engines, model optimization, GPU cloud computing and model architecture research.

Alongside its platform, Together AI contributes to open-source AI research and tooling. Its contributions include FlashAttention, an attention optimization for AI/ML tooling; Mamba, a contribution to AI model architectures; and RedPajama, a contribution to AI datasets and models. A commitment to open and transparent AI systems and shared knowledge shapes this work.

Engineers and researchers work side by side, with an emphasis on impact over hierarchy. Team members own substantial technical challenges from day one, and tackling frontier problems in distributed systems, model optimization and AI infrastructure is routine. The company describes a culture of deep technical curiosity, research-driven innovation and genuine collaboration, alongside its no lock-in approach to infrastructure.