Senior Product Manager · AI Tinkerer

Payments, product, and a healthy curiosity for AI.

I'm Anirudh. I spend my days on products millions of people use without thinking about them: at Bajaj Finserv, that's more than 15 million payments every month. Outside work, I build with AI: agents and automations that quietly handle the boring parts of my day.

AI agentsn8n automationsClaude & LLMsVibe-built tools

View experience LinkedIn GitHub Download résumé

Impact

The numbers behind the work

15.2M+

UPI payments every month, under my care

₹2Cr+

New revenue a year, from smarter offers

₹20Cr+

Purchases saved by better routing

7 yrs

Spent learning how money moves

A little about me

I've spent seven years on the quiet machinery behind products people rely on every day: the tap that pays, the card that gets approved, the offer that arrives at the right moment.

What I enjoy is the unglamorous middle: why a payment failed at 9pm, why a good customer was turned away, what a number in next year's plan really assumes. I like sitting with customers, testing rather than debating, and getting engineers, risk, compliance and finance to agree on one decision instead of four.

A few things I'm proud of

Outcomes I can point to

+10%

Ten percent more payments went through

Some payment routes were quietly failing more often than others. I found them, rebalanced how money was routed, proved it over a fortnight of live testing, and roughly ₹20Cr of purchases stopped falling through.

+₹2Cr

Offers that finally knew who they were talking to

Nearly a third of people were being shown things they couldn't even apply for. Teaching the system who to ask, and when, turned an irritation into more than ₹2Cr a year.

A misuse loophole closed for good

Around 110 people had been cheated through a gap in our auto-payment flows. I traced the pattern, tightened the controls, shipped the fixes, and the reports stopped entirely.

Where I've worked

Seven years of shipping

One employer-shaped career that turned out to contain a whole apprenticeship: credit cards, growth, and now the UPI rails themselves.

Nov 2022 – Present
Current
Senior Product Manager
Bajaj Finserv · Pune
I look after UPI and UPI Lite, the tap-to-pay rails millions of people use without ever thinking about them. My job is to make sure they simply work: signing up, paying, coming back.
That means launching UPI Lite and credit-card-on-UPI alongside the national payments body, sponsor banks and risk teams; forecasting a year of volume and revenue with Finance; and growing a small team across product, data and campaigns.
Jun 2021 – Oct 2022
Product Manager
Bajaj Finserv · Pune
I built the credit card sign-up journey for people entirely new to the bank, from the first tap to approval.
Leads used to arrive a day late; I got them flowing to the call centre in real time. Careful reading of where people dropped off made every marketing rupee go 65% further and cut the cost of each new customer by a third; around a third of monthly new cards came through this journey.
Aug 2020 – May 2021
Independent Product & Growth Consultant
Freelance
A year of helping small teams get found and grow, mostly on app stores and paid channels.
Moved a client's key search term from fifth to second place on the Play Store, which brought steady free downloads, and ran their advertising so each new sign-up cost less.
Apr 2019 – Jun 2020
Market Research Analyst
Hansa Research Group · Bangalore
Where I learned to ask people the right questions, and not trust messy answers.
Ran studies across food, consumer goods, property, clothing and public policy, including work for a national government think tank, and rebuilt how surveys were structured so the data could actually be relied on.

What I'm good at

The toolkit

Shaping the product

Deciding what to build and why: strategy, discovery, roadmaps, launches, and the honest experiments that settle arguments.

Product strategyDiscoveryRoadmappingPlatform productsGTMAgile

How money moves

The plumbing behind a payment: UPI, cards, the gateways in between, and everything that has to hold when volume spikes.

UPI & UPI LitePayment routingMerchant paymentsCredit cardsSettlement

Reading the numbers

Finding the real story in the funnel, then talking to customers to understand why it reads that way.

SQLFunnel analysisA/B testingSegmentationCustomer research

Building things myself

Enough hands-on craft to prototype an idea over a weekend instead of describing it in a document, and to wire up the boring parts so they run on their own.

Vanilla JavaScriptReactn8nClaude APITelegram botsNotion & Calendar APIsPostmanFigma

Things I built on my own time

Made on evenings and weekends

Built solo, in vanilla JavaScript

MVP live, still working out how to pull card rules automatically

Card Optimizer

Which card should you actually pay with?

India has 40-odd credit cards within arm's reach of any serious spender, and almost no honest way to compare them, so people leave cashback and points on the table every month. This tool asks how you spend, and answers which card leaves you better off. It ranks cards by net value (rewards earned minus the slice of the annual fee that month), shows the best card you already own, and quietly points at the one you should probably apply for.

Pick a category (groceries, flights, food delivery) and an amount
Every card scored on net value: rewards earned, fee impact applied
Side-by-side comparison across the whole wallet, recalculated live

40+ cards89 platforms10 categories500+ users

Vanilla JavaScriptZero dependencieslocalStorage

Try the live MVP

Built solo, with automation and AI

Paisapulse

An expense tracker with nothing to track.

I didn't want to hand my bank credentials, or read my messages, to a finance app, but I also wasn't going to log 400 transactions a month by hand. The insight was that every transaction already arrives in my inbox as a bank alert. So Paisapulse reads those emails, parses each bank's format with its own rules, removes the duplicates, and files everything into one living ledger. Claude sorts each spend into a category, and when I correct it, it remembers: it checks my corrections before spending a single API call next time.

Polls Gmail every 15 minutes, with bank-specific parsing and deduping
AI categorisation with a memory tab: my corrections stick, retroactively
A Telegram bot that answers in plain language, plus a dark live dashboard

400+ transactions/mo5 banks15+ categories

n8nGmail APIGoogle SheetsClaude APITelegram Bot APILooker Studio

Built solo, on Telegram

A second brain in my pocket

One chat window instead of five apps.

Capturing an idea, checking the news, blocking an hour on the calendar, each lived in a different app, and switching between them was where thoughts went to die. Now it all happens in one Telegram chat: a message arrives, its intent is detected, and it routes itself: ideas land neatly in Notion, natural language like "block tomorrow 3pm" becomes a real calendar event, headlines arrive summarised on request. The interesting problem wasn't the APIs; it was teaching an ambiguous message what it should actually become.

Forced intent detection, so "save this" becomes an action, not small talk
Writes to Notion over raw HTTP where the official node gave up
Conversation memory, so follow-ups understand context

5 apps replaced1 chat window

Telegram Bot APIn8nClaude APINotion APIGoogle Calendar APINewsAPI

Intent detection is a product problem, not just a technical one: the hardest part wasn't the API calls, it was defining what an ambiguous message should actually do.

Where I studied

Education

MBA

M.B.A

Christ University, Bangalore

Where I learnt to read a business the way a customer feels it: pricing, positioning and the numbers underneath. Most of what I do with revenue cases today started here.

2017 – 2019Marketing & Finance

B.E

B.Tech, Mechanical Engineering

Graphic Era University, Dehradun

Four years of taking systems apart to see why they fail. Payments turned out to be the same habit, with money moving through the pipes instead of heat and force.

2012 – 2016Systems thinking

If any of this sounds like your kind of problem, say hello.

Pune, Maharashtra · +91 91108 27955

aniguptm@gmail.com LinkedIn GitHub Download résumé