Policy-as-a-PDF → a live rules engine
Zero-config onboarding at Jisr: the customer uploads their expense policy, an LLM parses it into executable spend rules, and the approval workflow is generated from the employee graph.
10K+businesses on RazorpayX Corporate Cards, from zero — 10x growth two years running
₹500Cr+Loan Against Mutual Funds book in 9 months at Jio Finance, a first-in-market digital journey
6zero-to-one products led at Jisr, with teams across four countries
300%quarterly revenue growth from a new product line at Leena AI
How can
I help you?
Building in fintech, thinking about AI in a regulated product, or just here for the reading — pick a door.
Talk product & AI
Lending, cards, spend, or putting LLMs into a regulated workflow. 30 minutes, no deck needed.
Book a call →Read Context & Nuance
My Substack on all things product, growth and productivity.
Subscribe →Make sense of the RBI
The Plain Rupee: every RBI press release explained in plain words, with a daily digest.
Read the latest →AI-native track record
Building with AI, in production
AI in production, in a regulated domain
Shipped receipt OCR and the spend control engine — automating capture, policy checks and approval routing inside a payroll system where errors are audit findings, not bugs.
Conversational AI before it was a category
Head of Product at Leena AI (2018–20), one of India's first conversational-AI HR platforms; 300% quarterly revenue growth from an engagement product built on the bot.
An AI-augmented operating model
Discovery, competitive research, spec drafting and analytics run through LLM agents — Claude Code, MCP servers, local models, n8n/Composio — as the team's default, not a side experiment.
Depth to judge the claims, and the limits
M.Tech CS (Data Science) with a graph-ML thesis. A regulated background (RBI cards & digital lending, SAMA, WPS) means AI proposals get scoped against explainability, audit trail and consent up front.
Selected experience
Résumé (PDF) →
Where I've built
Saudi Arabia's leading HR and payroll platform. Owned the embedded-fintech portfolio for 3,000+ customers across four countries.
- Led six zero-to-one products — corporate cards, payroll wallet, earned wage access, expense, travel and insurance broking — lifting product attach rate by 20%.
- Conceived and shipped AI policy ingestion: an LLM compiles a customer's expense-policy document into executable spend rules and derives the approval chain from the employee graph, cutting enterprise onboarding setup to zero.
- Rebuilt the spend policy engine with warn-versus-block enforcement and branching approvals, shipped receipt OCR, and owned the build-versus-buy call on document AI.
- Built the payroll wallet so salaries disburse only from money the customer has deposited, satisfying WPS and SAMA rules; shaped the company's position on SAMA wallet limits, EMI and insurance-aggregation licensing.
- Replatformed travel booking from a partner iframe to a direct API — adopting companies grew from 35 to 300.
- Built the product org from scratch (3 PMs, 3 designers, a dedicated engineering group); portfolio OKRs cut time-to-market by 30%.
RazorpayX Capital — corporate cards, lending and settlements for Indian businesses.
- Scaled RazorpayX Corporate Cards from launch to 10,000+ businesses and ₹100Cr+ monthly spend, growing 10x in each of two consecutive years.
- Broke a 5–7% card-approval ceiling with a three-tier funnel: a risk-model limit, then a current-account-backed limit, then a zero-limit card on a credit BIN that still cleared ads spend.
- Designed the just-in-time authorization engine — approving or declining each transaction in 200–300ms, inside the card network's three-second deadline.
- Won the SBM Bank partnership on Razorpay-controlled underwriting through a 100% FLDG, with a capped risk-exception budget to prove the model before scaling.
- Stepped up to lead product for the Capital unit — corporate cards, cash advance and early settlements — growing SMEs and startups to 70% of the customer base and mentoring the PM team.
Travel and expense management SaaS — 4,500+ clients and 250,000 users.
- Built Happay V2, the enterprise line — expense, travel and accounts payable — winning Oxford University Press, PwC and Arangi Group against SAP Concur on configurability.
- Moved the product up-market with automated expense capture through ride feeds, corporate-card feeds and receipt OCR, cutting onboarding time by 50% and raising transaction creation by 19%.
- Shipped one of India's first combined corporate card and expense products, tying card spend directly to policy checks and reconciliation.
- Cut implementation and support tickets by 24%, and set up the product analytics behind a platform-wide navigation redesign.
Also
Jio Financial Services · Product Head, 2023 – 2024 — launched Loan Against Mutual Funds (₹500Cr+ in 9 months) and Loan Against Property & Home Loans (₹400Cr+ in 6 months)
Leena AI · Head of Product, 2018 – 2020 — conversational AI for HR; 300% quarterly revenue growth from a new engagement line
Uncvr (Evangelysis) · Cofounder & CEO, 2015 – 2017 — live-performance analytics; first 15 paying venues
Side projects
Things I've built lately
Shipped in the evenings with Claude Code and Vercel. Both live — click through.
The Plain Rupee
Every RBI press release explained in plain words — what changed, what it means for you and for businesses — with a searchable archive and a daily email digest.
Claude Code · Resend · Vercel
Tasca
The todo list that lives in Telegram. Type, forward or say a task — in Hinglish too — and it comes back exactly when it matters. No app, no account.
Telegram bot · Supabase · VercelKind words
What people say
The thing that will strike you about Mittul's product thinking is his uncanny ability of (un)black-boxing things — he can think from a 30,000-foot level as well as at a microscopic level of detail. During one discussion he went from his vision for independent music to the minute details of an algorithm that can potentially change how music recommendations work. I highly recommend Mittul, especially in the realm of Product Management.
It's grateful to work with a manager who is also a mentor to you. Mittul has always been a mentor to us and it's been a great learning curve working with him. He is communicative, listens to your ideas and ornaments them, which makes for a better idea. Working with him is learning a new thing every day.
He's very enthusiastic, energetic and utilizes logical solutions that can be practically implemented. I would recommend him for product ideation, business planning and business development projects.
Writing & talks
Substack newsletter
Context & Nuance
All things product, growth and productivity.
Subscribe
Podcast · ProductHood
Context & nuance
Best product managers are curious and read as much as possible across diverse fields. Degrees don't matter as long as you have that innate want to really learn.Listen: Conversations with Leena.AI's Product Head →
What I work with
AI & ML
LLM-native product designDocument AI / OCR pipelinesConversational AI & NLPAgentic workflows & MCP serversPrompt engineeringEval design & model selectionLocal LLM deploymentGraph MLClaude Code, n8n, Composio
Fintech
Corporate cardsDigital lending, secured & unsecuredInsurance broking & aggregationPayroll and walletsSpend management & T&ERBI (cards, digital lending, PPI/PA-PG), SAMA, WPS
Product
0→1 strategy & launchesPortfolio managementB2B SaaSRoadmap & prioritizationCustomer discoveryCompetitive researchOKRsGTM & pricing
Technical & leadership
SQLProduct analyticsA/B experimentationAPI designCross-functional leadership across four countriesHiring & team designBank, issuer and insurer partnerships
Foundations
IIT Hyderabad · 2014 – 2016
M.Tech, Computer Science (Data Science)
Thesis: COMBINED — Community Mining by Integrating Node and Edge Data; community detection on attributed-edge graphs via line-graph transforms. Head, Placement Cell · Head, Astronomy Club (Cepheid).
Gujarat Technological University · 2008 – 2012
B.E., Computer Science & Engineering
Head, Placement Cell · Vice Chair, IEEE and CSI Student Chapters.
Off the clock
A context chaser and nuance seeker — fond of reading, writing, music and design.