# Madhav Sharma

> Madhav Sharma builds the system and the audience: machine learning that runs in the browser, lead engines that find buyers, and growth from zero with no ad spend.

Madhav Sharma is Chief AI Officer at Quiny, an AI app builder based in New York, and works from Bhubaneswar, India. He builds technical systems, including a machine learning model that runs entirely in the browser, a lead engine that finds buyers in public data and a two-sided marketplace platform, and he grows audiences: he took a US startup’s new accounts to 10,000 followers in 10 days with no ad spend. He studies Data Science and Artificial Intelligence at IIT Madras.

Canonical site: https://madhavsharma.ai/
LinkedIn: https://www.linkedin.com/in/madhav-sharma-iitm/
GitHub: https://github.com/M4dhxv

## Key pages
- [About](https://madhavsharma.ai/about/): biography, principles and education
- [Experience](https://madhavsharma.ai/experience/): every role with dates, scope and outcomes
- [Work](https://madhavsharma.ai/work/): case studies
- [Contact](https://madhavsharma.ai/contact/)

## Case studies
- [Corporate Signal Translator](https://madhavsharma.ai/work/corporate-signal-translator/): A gesture classifier that runs entirely in your browser. The model and the runtime that executes it come to 1.7 MB, nothing is sent to a server, and a small decision layer makes it dependable.
- [Meela: 10,000 followers in 10 days](https://madhavsharma.ai/work/meela-growth/): I joined a US startup as a creator, ran its growth within six weeks, and took its new accounts from zero to 10,000 followers in 10 days with no ad spend.
- [Morpheus](https://madhavsharma.ai/work/morpheus/): A seven-stage pipeline that finds companies feeling a specific pain right now, from their job posts and public announcements, and hands sales a ranked list with the evidence quoted.
- [Raven](https://madhavsharma.ai/work/raven/): An agent for SEO and AI-search work that can only report what it can cite. Every finding has to point at a source a tool fetched in the same run, or it is rejected before it reaches the report.
- [BabyBrain](https://madhavsharma.ai/work/babybrain/): A two-sided marketplace for children's classes in Singapore, built from the first commit. Parents find and book, vendors manage listings, chat and payouts, and an LLM pipeline turns vendor websites into listings.
- [Short-form distribution systems](https://madhavsharma.ai/work/organic-distribution/): Faceless short-form pages built to reach people who had never heard of them. Over one 30-day window they reached 25.9 million views with no ad spend, 92 percent from new audiences.
- [Universal Financial Reporting Engine](https://madhavsharma.ai/work/financial-reporting-engine/): A reporting engine that takes any CSV or Excel export and produces a variance report with charts and a PDF, without being told what the columns mean.

## Topics
- [Signal-based lead intelligence](https://madhavsharma.ai/topics/signal-intelligence/): Finding the companies that are about to buy from what they are doing in public, such as hiring, restructuring and leadership changes, rather than from static firmographic filters.
- [Web data pipelines and LLM classification](https://madhavsharma.ai/topics/data-pipelines/): Collecting structured data from the open web at scale, then classifying and extracting from it with LLMs, so that a team can work from a clean table instead of a browser.
- [Reliable AI systems](https://madhavsharma.ai/topics/reliable-ai/): Making model output dependable enough to act on, with deterministic checks around the model, numbers computed in code, and a record of why each decision was made.
- [Organic distribution systems](https://madhavsharma.ai/topics/organic-distribution/): Treating short-form content as infrastructure for reaching cold audiences, with repeatable formats, creator programmes and automation from comment to conversation.

## Optional
- [Full text of every case study and article](https://madhavsharma.ai/llms-full.txt)
- [RSS](https://madhavsharma.ai/rss.xml)