Together AI

Together AI builds a purpose-built GPU cloud platform that lets developers and organisations of all sizes train, fine-tune and deploy generative AI models without vendor lock-in. The company's work spans several technical domains, among them generative AI, distributed systems, distributed inference engines, model optimization and AI infrastructure, alongside research into model architecture. Alongside its commercial platform, it maintains an active programme of open-source AI research, with contributions including FlashAttention, an attention-optimisation tool, Mamba, a model architecture, and RedPajama, a set of datasets and models.

Engineering work at the company is organised around problems in distributed systems, model optimization and AI infrastructure; the firm describes frontier problems in these areas as the norm rather than the exception. Engineers and researchers work side by side, and staff are expected to take ownership of substantial technical challenges from an early stage. The stated internal approach favours impact over hierarchy and is described as research-driven and collaborative.

Together AI positions its platform explicitly as free of lock-in, and applies a matching commitment to open and transparent AI systems and shared knowledge in its research output. The company operates in the artificial intelligence, generative AI and developer tools and cloud infrastructure sectors. Details on founding year, headcount, funding and geographic footprint are not published here.