What does it take to design the next generation of AI systems responsibly? In this episode of the Winners’ Circle Podcast, we sit down with ⁠Maneet Bansal⁠, a senior leader at Meta and longtime judge in our awards programs, to explore how observability, transparency, and trust are becoming essential in building AI at scale.

Maneet explains how AI workloads fundamentally differ from traditional cloud computing. In cloud environments, you can route requests across a pool of servers behind a load balancer. With AI, however, every request has unique demands on GPU memory, system resources, and downstream protections. It is not harder or easier. It is different. And that difference is reshaping the playbook for infrastructure and operations.

We discuss:

• Why observability is not just a tool but a philosophy for engineers and organizations that want to scale AI responsibly.
• The cybersecurity blind spots companies face when rushing to deploy large language models without guardrails.
• The importance of industry guidelines and collective responsibility to keep AI safe and trusted.
• How the speed of change in AI requires agility over long-term planning, with problems evolving faster than solutions.
• The role of mentorship, philosophy, and even poker in shaping Maneet’s approach to leadership and decision-making.

Maneet also shares two books that have shaped his thinking:

• ⁠The Courage to Be Disliked⁠ by Ichiro Kishimi and Fumitake Koga
• ⁠The Coming Wave⁠ by Mustafa Suleyman

Two standout quotes from the conversation:

• "It is not harder or easier. It is different." On the shift from cloud workloads to AI workloads.
• "The solution is not just smarter machines. It is clearer industry guidelines and collective responsibility." On AI and cybersecurity.

This episode is a reminder that the future of AI is not only about what is possible, but also about what is safe, transparent, and responsible.

Learn more about Maneet and his work as part of our judging community: ⁠Maneet Bansal Judge Profile⁠.