Nitesh Methani

Hi, I'm Nitesh.

I've spent 7+ years building NLP and AI systems—from creating PlotQA (28.9M question-answer pairs, one of the largest visual reasoning datasets) at IIT Madras, to shipping address intelligence at Amazon scale, to now building conversation AI agents at Sarvam as an early team member. But the credentials only tell part of the story.

I grew up in a Sindhi business family—the kind where you learn margins over dinner and watch your relatives renegotiate supply chains over chai before the banks even open. I saw shops reinvent themselves overnight when things broke, long before anyone in tech called it “pivoting.” That gave me something a CS degree alone doesn't: an instinct for what survives vs. what just demos well.

That instinct shapes how I build. At Amazon, I didn't just train a model—I dug into why deliveries were failing and built NLP systems that parse millions of Indian addresses where the “address” is often just a set of landmarks and local references. At Sarvam, I joined early because I recognized the pattern: a real problem (businesses drowning in unstructured call data), a clear path to revenue, and a team that ships. I also judge every product I build the way a first-time customer does—by how it feels, not how it works under the hood. If the UX isn't seamless, the engineering underneath doesn't matter.

And I'm a father—which changed how I evaluate what's worth building. I think in decades, not quarters. I want my kids to grow up in a world where AI makes people more capable, not more dependent—and where the builders of these systems cared enough to get the details right.

What I've built

Sarvam (2024–present): Early team member building Agent & Conversation Analytics—turning raw call transcripts into actionable intelligence. Every conversation makes the system smarter.

Amazon (2020–2024): NLP and deep learning for address parsing at scale across India. Published in Amazon Science.

IIT Madras (2018–2020): Created PlotQA—28.9M question-answer pairs over 224,000+ scientific plots. Worked under Mitesh Khapra and Pratyush Kumar at RBCDSAI.

Tayari (2025–present): Every exam prep app I found was either paywalled, ad-riddled, or trying to upsell a ₹50K coaching package. So I built Tayari—a self-study, data-driven, AI-powered platform with 15+ years of topic-wise PYQs, 2,200+ flashcards, spaced repetition—all in a distraction-free interface. Born from my own GATE experience. 2,000+ students use it, and they keep calling it “buttery smooth.”

What I'm exploring

Reverse-engineering multi-agent systems like Claude Code and Cursor. Reading about consumer behavior and neuromarketing. Building mobile apps—from a Gandalf-voiced history narrator to conversational language-learning chatbots.

Get in touch: initeshmethani@gmail.com · Twitter