About

I’m an AI Engineering Lead and a builder who likes owning hard problems end-to-end — especially when the path from “this might be possible” to “this is running in production” isn't obvious.

Over the last 6+ years, I’ve worked across Agentic AI, Generative AI, machine learning, computer vision, and large-scale AI systems, turning ambiguous problems into products that deliver measurable business impact.

At Ksolves, I lead a 20-member AI team building solutions across GenAI, Agentic AI, and traditional ML. One of the systems we built replaced manual support workflows with autonomous AI agents and now resolves 70% of support tickets without human intervention. My role spans architecture, engineering decisions, and taking these systems all the way to production.

Before that, at Paytm, I took on the problem of scaling merchant support with AI. I built and shipped a conversational AI system end-to-end using LLMs and in-house NLP models that ultimately replaced 25% of merchant help-desk agents and reduced cost per call by 15×.

I also built a lender recommendation system that reduced sales-funnel dropout by 87%, owning everything from feature engineering to the production API.

Earlier at Infogain, I architected AI & A-Eyes, a multimodal computer-vision and NLP platform for retail monitoring. What started as an AI engineering problem became a production platform generating $14 million in annual revenue and processing roughly 13 GB of image data every day on Azure.

And even earlier, I helped build an AI-powered learning platform capable of evaluating handwritten student assignments using computer vision and NLP — technology that ultimately reached 1,500+ schools across India.

Along the way, I’ve been named Employee of the Year, received multiple Employee of the Quarter awards, represented Ksolves at the Gartner Application Summit in Las Vegas, presented research, and received The Indian Express' Most Innovative Use of AI award.

But titles and awards aren't really what keep me interested. I like building things. I like understanding systems deeply enough to simplify them, challenging abstractions instead of blindly using them, and turning what I learn into something other people can actually use.

This website is an extension of that mindset — my public visual knowledge space, where I break down AI, engineering, architectures, systems, and whatever else I’m exploring into interactive boards.