AI Product Manager. 8+ years turning messy enterprise workflows into agents, APIs and search that teams actually adopt.
I sit where AI systems meet real business problems, and I don't stop at the spec. I prototype, test and ship alongside the engineers.
I started as a software engineer, moved into product strategy at Aditya Birla Capital and ZS Associates, and now lead AI platform products at Prezent AI. That path lets me go deep on architecture with engineers, frame trade-offs for executives, and sit with users to see how the work really happens.
Outside work I build fintech and AI-trust prototypes end to end, each with a PRD, a decision log and numbers you can check.
Agents, RAG, semantic search and MCP, shipped with evals and guardrails.
API-first design, OAuth, webhooks and integrations enterprises build on.
Recommendation, incentive automation, fraud and credit decisioning.
Engineering, Data Science, Design, Ops and CXOs on one roadmap.
Twelve products across Prezent AI, ZS Associates and Aditya Birla Capital. Open any one for the problem, the approach and my role.
Personal projects where I own everything: problem framing, PRD, prototype, metrics and write-up.
Live
A no-code simulator where payment risk teams back-test fraud rules on last month's transactions before going live, so every change shows fraud caught vs good customers blocked, priced in rupees.
Live
AI tags youth-sports clips to each player's reel, but look-alike jersey numbers put clips in the wrong kid's reel. I designed a confidence gate: a one-tap coach confirm for doubtful clips. Anything unconfirmed is left out, never guessed.
Credit policy rules, scorecards and approval vs default trade-offs.
Interchange, credit losses, cost of funds and CAC.
Discovery to launch: strategy, design, engineering collaboration, data and AI, plus the leadership that holds it together.
I build working prototypes in Claude Code and Lovable to validate with stakeholders before engineering invests, cutting idea-to-validation from sprints to days.
Discovery and research on how the work actually happens.
Working prototypes validated with users and stakeholders.
Architecture, data model and human-in-the-loop, agreed up front.
Evals, failure-mode analysis and human oversight at launch.
Activation, adoption and retention KPIs, A/B tests, OKRs.
Great Lakes Institute of Management, Gurgaon
Guru Nanak Dev University, Amritsar