Product Manager · Fintech · Insurtech · GenAI · Bengaluru
I find the problems worth solving — then build the case to solve them. AI-powered KYC for tier-1 Indian banks, a $10M GenAI transformation for US healthcare, and roadmap strategy for an insurance platform at 500K+ customers.
From e-commerce to fintech to enterprise AI — each role a different PM discipline.
Products I've defined, designed, and documented from scratch — showing how I think when there's no brief and no backlog.
Job seekers had raw data. They had no signal. I designed a product that extracts structured hiring intelligence — skill demand trends, company hiring velocity, salary signals — from unstructured job posting text. The key insight: prompt engineering is acceptance criteria. The precision required to get consistent AI output is identical to the discipline of writing good feature specs.
Early-stage founders lose 40% of their week to tool-switching. I defined a zero-to-one product strategy for a unified decision-making layer — connecting OKR tracking, investor CRM, hiring pipeline, roadmap, and burn rate. The core PM question: what is the minimum slice of functionality that creates irreversible founder habit?
End-to-end narratives — discovery, prioritization, the bet I made, and what the data said after. Not what I did, but how I decided.
Banks told us onboarding took 3 days because of slow processing. Digging into 3 months of transaction logs revealed the real problem: 60% of users dropped in the first 10 minutes with no feedback. Switching north star metrics from "time to approve" to "first-attempt success rate" changed the entire roadmap — and saved 8 weeks of misdirected engineering.
$10M program. 2,000+ agents. A client who wanted "AI in the contact center" but couldn't define what that meant. I built a risk-adjusted prioritization framework and argued against shipping the feature everyone wanted — in order to ship the one that would earn us the trust to build what actually mattered.
I inherited a backlog where every client claimed P0, engineering had 6 release slots, and the math didn't add up. Built a 4-factor prioritization framework (client breadth, revenue linkage, strategic value, technical risk) that made every tradeoff legible and defensible — to clients, to engineering, to leadership.
Every client said "fix our checkout." Checkout was fine. Session recordings showed users were abandoning at the product detail page — they couldn't find what they wanted. Betting on a personalization engine instead of a checkout optimization was the counterintuitive call that drove +16pp conversion across three premium brand accounts.
Product teardowns, strategy documents, PRDs, and frameworks — the PM thinking that lives above the roadmap.
Business model deconstruction, feature analysis, competitive positioning, and 3 product opportunities I'd build next.
How a developer-first payments API became India's SME banking infrastructure — and the PM questions behind every platform bet they've made since 2014.
Live experiments, frameworks built from real programs, the AI product checklist I run before every feature ships, and what I'm watching in AI, India fintech, and agentic systems. A working document — updated as I go.
Competitive positioning of J.P. Morgan Chase vs. Goldman Sachs, Morgan Stanley, and Citi across digital banking. SWOT and QSPM analysis leading to a $1.5B 3-year product investment roadmap for emerging market expansion. The PM question at the center: where should a bank that has everything focus its digital product bets?
Four PRDs showing how I write specs that engineers actually build from: JobSignals (AI data product), KYC Guided Capture (RBI compliance-annotated), GenAI Email Drafting (adoption-first design), and Founder OS (zero-to-one). Each with a "signature decision" — the most interesting tradeoff in the spec.
Three strategic documents: the AI Feature Risk Prioritization Framework (built from the HCSC program), an India Fintech Product Strategy Brief (3 bets for 2025–2027), and an Enterprise Insurance Platform 18-Month Strategy. The PM thinking that lives above the roadmap.
A practical guide to building AI-powered products — moving past prompt engineering to real product thinking, architecture decisions, and shipping AI features that stick.
AI product management, fintech strategy, and the craft of building products in regulated industries.
How measuring "time to approve" instead of "first-attempt success rate" sent our KYC product team in the wrong direction — and what switching metrics changed about the entire roadmap.
Read article arrow_forwardThe client wanted AI in the contact center. The most-requested feature had the highest failure cost. Here's the risk-adjusted framework I used to sequence 10+ GenAI features — and why I argued against the headline feature in Sprint 1.
Read article arrow_forwardThe most technically impressive AI feature we shipped had the lowest adoption. The fix wasn't a better model. It was one design decision — and the skeptical users who refused to trust the AI turned out to be our best model trainers.
Read article arrow_forward
Everything a tech-curious person needs to know about LLMs, RAG, agents, and building real AI products — without the fluff.
Read on Medium open_in_new
Why the hardest part of enterprise fintech integrations isn't the code — it's the spec. How configuration dependencies and schema misalignment cause production failures, and what PMs can do about it before engineering starts.
Read on Medium open_in_new
First-person account of implementing API integration strategy in banking — cutting onboarding 40% and improving KYC accuracy 35%.
Read on Medium open_in_new
Structured framework for user stories that engineering teams actually want to read — capability classification, design integration, acceptance criteria.
Read on Medium open_in_new
A walkthrough of the discovery process that turns chaotic operational workflows into a clear product brief — using a restaurant OMS as the case study. The PM lesson: always map the current state before proposing a solution.
Read on Medium open_in_new
There's a difference between a PM who uses data and one who understands it. How data fluency — not just dashboards — changes the quality of product decisions you're able to make at speed.
Read on Medium open_in_newComplete PRD template for integrating Razorpay into e-commerce platforms — problem space, solution design, integration steps, and success metrics.
Read on Medium open_in_newMost PM thinking dies in Confluence docs nobody reads. I write to force the clarity I need before I can make a product call — and to stress-test beliefs I'm not sure I can defend yet. If I can't write it clearly, I can't build it confidently.
Follow on Medium open_in_newThe person behind the portfolio — what drives me, what I'm working on, and how I think.
Most product failures happen before a single line of code is written. They happen when the team agrees to solve the wrong problem. My job is to prevent that. Six years in KYC, healthcare AI, and enterprise insurance taught me that the hardest product questions are never technical — they're about which problem is actually worth solving.
At Iksula, every client said "checkout is broken." Checkout was fine. Discovery was broken. I learned to be professionally skeptical of the first diagnosis.
At Signzy, we were tracking "time to approve." The right metric was "first-attempt success rate." Switching metrics changed the entire roadmap — and saved 8 weeks of misdirected engineering effort.
A healthcare AI right 90% of the time but catastrophically wrong 10% is worse than no AI. I sequence AI features by asymmetry of failure, not by impressiveness.
At HCSC, the most technically impressive AI feature had the lowest adoption — until we redesigned the UX for the skeptic, not the enthusiast. Feature quality ≠ feature adoption.
Enterprise clients who trust your release cadence take on more integration risk and expand faster. Every clean release is a trust deposit. I track this as deliberately as any feature metric.
From engineers, programme managers, and stakeholders across four organizations.
Shraddha has a rare quality: she asks the uncomfortable question — 'are we solving the right problem?' — before anyone else in the room does. On our HCSC programme, she was the one who stopped us from shipping the wrong feature first. That call saved the programme's credibility in Phase 2.
What I valued most about working with Shraddha was how precise her specs were. As an engineer, a clear spec is the best thing a PM can give you. She never wrote requirements by assumption — every edge case was documented, every dependency was mapped. I never had to guess what she meant.
Shraddha was the first PM I'd worked with who could articulate the downside risk of a feature before I asked. That level of thinking made it easy to trust her recommendations — even when she pushed back on what we thought we wanted. She made our team better at asking the right questions.
Full recommendations available on LinkedIn
Business strategy meets engineering depth — the combination that shapes how I think about products.
Structured learning that reinforces the judgment behind the work.
Validated ability to manage product backlogs, define and communicate product value, and work effectively with Scrum teams in iterative delivery.
check_circleActiveCloud fundamentals — compute, storage, databases, IAM, and core AWS services. Bridges PM-to-engineering conversations on infrastructure and scalability decisions.
check_circleActiveStructured problem-solving, communicating under uncertainty, and digital strategy frameworks from McKinsey's flagship leadership and capability-building program.
task_altCompletedLooking for PM, TPM, and AI PM roles in Fintech, Insurtech, and AI-first product companies. Open to remote-first teams globally. I respond to all PM role inquiries within 24 hours.