Skip to content
View jayvadolkar's full-sized avatar
🚩
Peace
🚩
Peace

Block or report jayvadolkar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jayvadolkar/README.md

Hey, I'm Jay 👋

LinkedIn Portfolio Email

📍 Bengaluru, India · Product @ Zoop.One · Ex-Keka, Manzuri · GrowthX Fellow (GX14)


The short version

I'm a Product Manager with 3+ years building AI-first products across InsureTech and HRTech. I started out writing code, and I never fully put it down — I still spec, prototype, and ship alongside engineering. That's the lens I bring: I'd rather build the thin slice and watch a real user touch it than argue about it in a doc.

What I actually care about is the gap between a messy real-world problem and the smallest product that closes it. Lately that gap is full of LLMs, agents, and APIs — which is exactly where I like to be.

Right now: building GenAI claim processing at Zoop.One — taking insurance claim approvals from 5 days down to 12 hours across 6,000+ claims a month.


🧪 Selected work

Each one is a problem I owned end to end — here's the bet I made and what happened.

🛡️ Zoop.One — GenAI claim processing for insurers  (2026 → now)

The problem. Insurance claims were stuck in manual review — five days to approve, reviewers buried in documents, drop-off everywhere, and IRDAI compliance leaving no room for error.

The bet. Don't replace the reviewer — un-bury them. Build an LLM + vision document-intelligence layer that reads, extracts, and triages every claim before a human looks at it, then design onboarding tight enough to survive regulation.

What happened
  • ⏱️ Approval TAT cut from 5 days → 12 hours across 6,000+ claims/month
  • 🤖 LLM + vision document-intelligence module eliminated manual review, with a 4% fraud-flag rate catching what humans missed
  • 🚀 Designed the end-to-end RO activation flow — onboarding 10+ insurer regional offices from credentialing to first live claim
  • 🎯 Built segment-specific onboarding across 4 insurer types to cut funnel drop-off, all within IRDAI compliance
🧩 Keka — self-serve developer portal & marketplace  (2023 → 2025)

The problem. Partners wanted to build on Keka, but integration took 60 days of hand-holding. There was no self-serve path from "I have an API key" to "I'm live in the marketplace" — and PMs were drowning in support triage along the way.

The bet. Turn a high-touch services motion into a product. Ship an App Builder + developer docs + marketplace that takes a partner from app creation → permission scoping → webhooks → sandbox → live listing without a human in the loop. Then point AI agents at the busywork.

What happened
  • 🤝 20+ partners taken from zero to live
  • ⚡ Partner integration time cut from 60 days → 15 days
  • 💰 Drove API–product parity from 35% → 75%, fueling 40% growth in API product revenue
  • 🛒 Launched the in-product marketplace — partner revenue +20% MoM, scaling partner + API-led MRR to ₹1 Cr
  • 🧠 Built an agentic ticket-triage system (local LLMs + Claude API) that cut PM workload 30%
  • 🔌 Shipped a Keka API MCP server so coding agents could integrate against Keka's APIs with baked-in flows
App Builder · Developer Docs · Marketplace · OAuth 2.0 · 20+ webhook events with HMAC verification · sandbox + retry logs

📄 Full case study →

📚 Product teardowns & assignments

Self-driven case studies — how I scope a problem, find the leverage, and design the bet.


🧭 How I think about product

  • Outcomes over output. A shipped feature that doesn't move a number is just code I have to maintain.
  • The roadmap is a hypothesis. It stays one until a real user proves it. So I talk to users early and often.
  • Build the thin slice. Prototyping is the argument — I'd rather show than debate.
  • AI is a means, not the pitch. I reach for LLMs and agents when they collapse a workflow, not because they're on the slide.

💼 Experience

Keka HR  ·  Product — AI & Marketplace Integrations  ·  May 2023 – Nov 2025
Grew from Product Analyst to APM on the platform team — built the developer portal, marketplace, and agentic tooling above.

Manzuri  ·  Product Analyst  ·  Oct 2022 – May 2023
Owned D2C growth for a stigmatized category: +30% engagement, +12% MoM revenue, +27% CSAT, −30% churn.

GrowthX®  ·  Fellow, GX14  ·  Mar 2023 – Present
Product & Growth program — graduated with a top-5 team.


🛠 Toolkit

AI & ML  ·  MCP (Model Context Protocol) · LLMs (GPT, Claude, OpenAI, LangChain) · AI Agents · RAG · Fine-tuning · Prompt Engineering · n8n · Lovable · bolt.new

Product  ·  Onboarding & Activation · Funnel Optimization · Developer Portals · API Ecosystems · A/B Testing · PRD Writing · Roadmap Management · Customer Discovery

Technical & Analytical  ·  Python · SQL · Power BI · JIRA · Postman · Azure DevOps · Confluence · Mixpanel · Pendo · API Integrations

Claude MCP Python SQL Jira Postman Mixpanel Power BI Figma


🎓 Education & honors

  • B.Tech, Computer Science (Gaming Tech) — Vellore Institute of Technology · 2019–2023
  • Product & Growth Program (Top 5 Team) — GrowthX® · 2023
  • 🏆 Economic Times Campus Star

✉️ Let's build something

I'm always up for a conversation about AI products, developer platforms, or a good 0→1 problem.

LinkedIn Portfolio Email

Interests: Product Strategy · Onboarding · Data-Driven Growth · Risk & Compliance · Behavioral Analytics · AI in Product · FinTech & RegTech

Popular repositories Loading

  1. NavMesh-Tutorial NavMesh-Tutorial Public

    Forked from Brackeys/NavMesh-Tutorial

    Tutorial project files on using NavMesh in Unity.

    C#

  2. 2048 2048 Public

    2048 Game

    C#

  3. sci-calculator sci-calculator Public

    Scientific Calculator using Python

    Python

  4. jayvadolkar jayvadolkar Public

    Config files for my GitHub profile.

  5. coursera coursera Public

    CSS

  6. lab-cricpro-scoreboard lab-cricpro-scoreboard Public

    Forked from prograd-org/lab-cricpro-scoreboard

    HTML