This week, we’re spotlighting Lele Cao, a Senior Principal AI Researcher and Research Lead at King AI Labs, Microsoft Gaming, who translates cutting-edge machine learning research into practical impact for large-scale mobile games. With a unique background rooted in clinical neuroscience, Lele’s journey into AI was sparked by the biological thresholds of neurons-an inspiration that now drives his work in representation learning, time series modeling, and graph neural networks.
Beyond making games smarter and more fun, Lele is unlocking the “LLM moment” for complex event streams and co-founding the science verification platform Scholar7. In this profile, he shares his perspectives on the evolution of noisy data, his reliance on AI-powered co-pilots, and why the future of the field belongs to those who anchor deep technical expertise in real-world applications.
Hyperight.com: What’s the best way to describe your job to someone outside tech?
Lele Cao: I teach computers to understand player behavior at scale so that games can become smarter and more fun.
Hyperight.com: What originally sparked your interest in AI/data, and what keeps you inspired today?
Lele Cao: My early clinical education in neuroscience showed me how neurons integrate inputs and fire only when a membrane threshold is crossed – an idea that drew me to neural networks, where activation functions similarly enable nonlinear intelligence to emerge. What keeps me inspired today is that AI is no longer a niche academic pursuit but a societal-scale transformation; we’re collectively witnessing a technological shift that is reshaping how we work, create, and think, and it’s both exciting and humbling to contribute to that revolution.
Hyperight.com: What is one challenge you’re trying to solve, and why does it matter?
Lele Cao: Learning meaningful representations from large-scale, irregular, and dynamic event streams. In gaming and many other industries, data is continuous, noisy, and constantly evolving. If we can model this complexity effectively, we essentially unlock the “LLM moment” for even streams. This matters because better representations lead to better decisions – whether in games, finance, healthcare, or beyond.
Hyperight.com: A tool you can’t live without (tech or not)?
Lele Cao: Coding co-pilots (such as Codex) and Paper co-pilots (such as CSPaper.org).
Hyperight.com: What trend in data or AI do you think will shape the Nordic region the most?
Lele Cao: Applied foundation models, especially those adapted to domain-specific and industrial contexts. The Nordics have strong engineering cultures and high digital maturity, which create ideal conditions for translating AI research into responsible, scalable products.
Hyperight.com: What’s one piece of advice you’d give to others entering the data and AI field?
Lele Cao: Build strong fundamentals in mathematics and modeling, but don’t stay in theory alone. Work on real problems, collaborate across disciplines, and aim for impact. The future belongs to those who can connect deep technical expertise with meaningful applications.
