作者:TOI World Desk / TIMESOFINDIA.COM / Updated: Aug 05, 2025, 13:31 IST
Andrew Tulloch, an Australian computer scientist and machine learning expert, has made headlines after turning down a $1.5 billion offer from Mark Zuckerberg to rejoin Meta. A former OpenAI engineer, Tulloch is now the co-founder of Thinking Machines Lab alongside ex-OpenAI CTO Mira Murati. The AI startup, still in its early stages, is already valued at $12 billion. Tullochâs decision to decline Zuckerbergâs aggressive recruitment attempt reflects a broader trend of top AI talent prioritizing independence, mission-driven work, and long-term impact over staggering financial packages.
Tullochâs journey began at the University of Sydney, where he graduated with first-class honours and a University Medal in mathematics. He later earned a Masterâs in Mathematical Statistics from Cambridge and pursued a PhD at UC Berkeley. Tulloch worked at Meta (then Facebook) from 2012 to 2023, contributing to machine learning systems and the development of PyTorch. He played a pivotal role in making PyTorch one of the most widely adopted frameworks in AI. His impact was felt across Metaâs core AI infrastructure. He joined OpenAI in 2023, focusing on GPT-4 pretraining and reasoning models, before co-founding Thinking Machines Lab in early 2025.
Unlike many AI engineers who stay within big tech, Tulloch has consistently sought out roles where he can shape core AI architecture rather than optimize existing systems.Co-founded with Mira Murati, Thinking Machines Lab is focused on building AI systems that are safer, interpretable, and customizableâgoing beyond traditional chatbot interfaces. The startup, though yet to release a product, has secured a $2 billion seed round with backing from Andreessen Horowitz, Nvidia, AMD, and Google Cloud. Its ambition and leadership have made it a top target for recruitment, especially from Zuckerberg's new "superintelligence" division at Meta. Industry insiders say the startup has quietly attracted talent from DeepMind, Anthropic, and OpenAIâfurther signaling that it may become a major challenger in the foundation model race.
According to The Wall Street Journal, Zuckerberg personally tried to lure Tulloch back to Meta with a six-year offer worth up to $1.5 billion, contingent on bonuses and stock performance. Tulloch declined, joining other Thinking Machines Lab co-founders in resisting Metaâs poaching attempts. His decision has sparked widespread praise among AI researchers, many of whom see it as a stand for academic and scientific freedom. The bold rejection has gone viral, with Tullochâs LinkedIn profile celebrated for charting a rare career driven by principles rather than payouts. Meta later disputed the exact terms of the offer but confirmed outreach efforts were made.