A benchmark isn't a win
A model only matters when it holds up in production, in front of real users. That's the bar I build to - not the leaderboard.
From training NLP models in a lab to co-founding an AI company - here's the path, and what drives it.
I'm a machine learning engineer and researcher with 5+ years building AI systems - specializing in NLP and large language models. Today I'm Co-Founder & Head of AI at Purands, building agents that plan and act on their own.
From training NLP models as an intern to leading AI at a company I co-founded, the throughline has stayed the same: build systems that don't just predict - they decide, and act. My north star is to make AI that meaningfully amplifies what people can do, steadily toward Artificial General Intelligence.
My first taste of turning data into insight - personality prediction for Samsung users from their MyGalaxy data.
Led in-house NLP models across hundreds of domains - intent classification, entity extraction, summarization, QA, and ASR.
Built competent AI agents capable of autonomous planning and action, driving advancements toward AGI.
Building the agentic retention suite from the ground up - and the AI engine at its core.
Build AI that doesn't just predict - it decides, and acts.
A model only matters when it holds up in production, in front of real users. That's the bar I build to - not the leaderboard.
The shortest path to a great system is a working one shipped early, then sharpened every week with what users actually do.
Every project is a step toward autonomous, genuinely useful AI. I'd rather make the move that compounds than the one that demos.
Integrated M.Sc. (Honours), Mathematics & Computing · 2015 - 2020
See what I'm building at Purands and the work that led here.
Explore my work →