CTO |Thinking Machines Lab
Co-creator and long-time lead of PyTorch. Spent 11 years at Meta FAIR as VP of AI Infrastructure. Co-authored foundational GAN papers (DCGAN, Wasserstein GAN). Now CTO at Mira Murati's Thinking Machines Lab.
Biography
Soumith Chintala is an Indian-American AI researcher and engineer best known as the co-creator and long-time lead of PyTorch, the open-source deep learning framework that powers the majority of the world's AI research and production systems. He spent eleven years at Meta (formerly Facebook) AI Research (FAIR), where he rose to VP of AI Infrastructure and contributed to foundational work on GANs, StarCraft AI, and large-scale computing clusters. After leaving Meta in November 2025, he was appointed CTO of Thinking Machines Lab, the AI startup co-founded by former OpenAI executive Mira Murati. Chintala studied at Vellore Institute of Technology (VIT) in India before earning his MS in Computer Science at NYU under the mentorship of Yann LeCun.
Co-created and led the PyTorch deep learning framework for nearly eight years, growing it from a research tool to the dominant platform powering AI research and production at hundreds of companies, research labs, and educational institutions worldwide.
Co-authored the seminal paper introducing Wasserstein distance as a training objective for GANs, solving mode collapse and training instability issues. Published at ICML 2017.
Co-authored the paper on unsupervised representation learning with deep convolutional GANs, one of the most cited works in generative modeling, establishing architectural guidelines for stable GAN training.
Built the first FAIR computing cluster and contributed to scaling Meta's AI infrastructure to 600K+ H100-equivalent GPUs, enabling training of Llama and other foundation models.
Championed open-source AI at Meta with the philosophy 'if it helps the world, it probably helps us,' answering thousands of community questions on PyTorch and Torch forums and advocating for a centralized feedback system for open-source models.
I needed to do something small again. I couldn't live with the counterfactual regret of never trying something outside Meta.
I probably have one of the AI industry's most leveraged seats... This kind of power is really hard to give up.
Closed AI rate-limits against people's imaginations and needs.
I grew up in India... knowledge was centralized, but open source helped me learn faster for zero dollars.
I don't want to be like Guido or Linus -- bound to a single thing for decades.
Keeping close to the grassroots is really important to build great products and organizations.
Research generated March 19, 2026