Saran Krishnasamy

I’m the co-founder and CTO at Gigit.ai where we’re building AI-native personalization for ecommerce brands. We don’t build dashboards, we build live dynamic storefronts, conversational product discovery and real-time recommendation that drives conversion rates. Prior to GigitAI, I was a staff AI engineer at Dataminr for almost 10 years, where I built real-time event detection systems for processing huge event volumes with low latency. That work has informed my thinking on the gap between research-grade ML and production ML.

I publish in NLP and IR venues such as ACL, LREC and most recently a SIGIR 2026 paper on constraint-aware conversational search, and peer review for top-tier NLP and AI conferences. I have a Master's degree in CS from Cornell.
What I care about is the seam between retrieval, large language models and real user intent. That's where most of the interesting problems in applied AI live right now: how to surface the right answer fast, with calibrated confidence, under real latency and cost budgets. That’s the common thread running through my research, my work at GigitAI, and the way I think about where the field is headed.