Shubho Sengupta
Deep learning researcher building a self-improving AI mathematician at Axiom Math.
Shubho Sengupta builds self-improving AI systems. He is currently the CTO of Axiom Math, where he is working to create an AI mathematician. Before this, he spent a decade in deep learning research at Meta and Baidu.
His past work covers speech models, reinforcement learning for Go, and foundational GPU programming. Today, his focus is on LLMs and automated software systems. He speaks about the realities of trusting machine-written code, exploring how engineering teams can guarantee safety when human review falls short.