Sumit
Ranjan

I'm Sumit Ranjan — currently in stealth, focused on a single hard problem: making agentic AI systems reliable, secure, and cheaper to run. That means building compact Small Language Models (SLMs) that are domain-trained, verifiable, and enterprise-ready — without the cost and opacity of large general-purpose models.

My research at BITS Pilani (PhD) sits at the intersection of agentic system architecture and AI security — exploring how autonomous agents fail, how they can be exploited, and how to build them so they don't. I contribute to the OWASP Agentic AI Top 10, helping define the security standards the industry will build upon.

I'm also the author of Enterprise Agentic AI (Packt) — a practical guide to deploying autonomous AI in production. As a Responsible AI Leader and keynote speaker, I bring this research into boardrooms, conferences, and classrooms globally.