I turn fuzzy problems into shipped products with measurable outcomes — and I build with data and AI as the moat, not the garnish. 5+ yrs · 0→1 to enterprise · AI/ML, Insurtech, Healthtech, Consumer & B2B SaaS.
55% → 93% Indic LLM accuracy · +15pp D1 retention @ 150M+ users · ~$5/claim agentic insurance pipeline
Spotify Premium 5%→12.5% (20x ROI) · Lenskart ₹12,600 Cr model (116x ROI) — and more in the case studies below.
🚀 New here? Pick one. Each card is a doorway — open the one that matches what you're hiring for.
| 🧠 AI & Data Portfolio — production AI shipped across B2C, D2C, B2B, SaaS, IaaS, AI/ML, B2B2C & Enterprise. Start here if you're hiring for AI. | 📂 Case Studies — 8 end-to-end products, problem → measured outcome. Start here for product depth. |
| 📈 Metrics & Experiments — North Stars, metric trees, experiment design. Start here for rigor. | 👤 About Me — how I work, what I believe, what I'm looking for. Start here to know the human. |
| 🔬 Problem-first depth | 🤖 Production AI, not demos | 📈 Metrics-driven | 🧩 0→1 systems thinking |
|---|---|---|---|
| First-principles discovery — 5W1H, 60+ interviews, 15K+ reviews mined — before any solution. | Shipped an Indic LLM 55%→93%; multi-agent claims at ~$5/claim. | North Star + metric tree + experiment on every product. | Founding PM of MyUni; map states, flows, feedback loops & trade-offs. |
The range that matters: I've applied AI and data thinking from consumer scale to regulated enterprise. One profile, eight contexts.
| Business model | Where I've shipped / designed it | AI / Data role |
|---|---|---|
| B2C / Consumer | ShareChat·Moj (150M+), Spotify | Vernacular ranking, conversion funnels |
| D2C / Retail | Lenskart, SastaSundar | Re-engagement loops, OCR extraction |
| B2B / Enterprise | Agentic Insurance Claims | Multi-agent pipeline, HITL, explainability |
| B2B SaaS | Cheerio.ai | Agentic commerce, outcome-based pricing |
| AI/ML · platform | AI4Bharat / IIT Madras | LLM eval, data pipelines, 22 languages |
| B2B2C / marketplace | LinkedIn Verified Badge, MyUni | Trust layers, cold-start, network effects |
→ Full breakdown in the AI & Data Portfolio.
Eight products, one consistent spine: problem → users → hypothesis → execution → measured outcome.
| Case Study | Domain | Model | Key Outcome |
|---|---|---|---|
| MyUni — Trust-Layered Student Community | EdTech · Community | B2C / B2B2C | Founding-PM 0→1: discovery → pricing → PMF → GTM |
| Agentic Insurance Claims — Document Processing | Insurtech | B2B / Enterprise | Multi-agent pipeline at ~$5/claim |
| Cheerio.ai — B2B Pivot to Agentic Commerce | B2B SaaS | B2B SaaS | Outcome-based pricing; $40M ARR forecast |
| SastaSundar — Pharmacy Optimization | Healthtech | B2C / D2C | OCR + overnight-queue fix on GMV |
| Lenskart — Family Vision Vault | Retail D2C | D2C | ₹12,600 Cr model, 116x ROI |
| Spotify — Premium Conversion | Consumer | B2C | 5% → 12.5% conversion, 20x ROI |
| LinkedIn Verified Badge — India-First Adoption | Marketplace | B2B2C | Cold-start trust & adoption strategy |