The Tiebreaker
A decision assistant. It structures a choice into pros and cons, comparison tables, or a SWOT, so the person can see the reasoning and make the call.
live demo ↗ · code ↗Monika Damor.
Product Manager·Carnegie Mellon ’26·San Francisco, CA
Growth PM. 6 years across acquisition, retention, and engagement.
Ask first, scroll later.
06
Years in product & growth
Fintech · Commerce · Logistics
350M+
The scale I’ve worked at
Paytm · user lifecycle
07
Prototypes built this year
Public on GitHub
03
AI certifications
Google · Maryland · Pendo
01 · About
About me.
I’m Monika, a growth product manager. For six years I owned onboarding, activation, and retention on consumer fintech and commerce products, at Paytm (payments), ShopX (B2B commerce), and two early-stage startups in logistics and home healthcare. I finished a master’s at Carnegie Mellon University in May 2026.
What I’m good at is the loop between data and decisions: run the experiment, read the funnel, write the one-page doc that gets leadership to a call. This past year I added a step to that loop. With AI I can build the thing itself, not just the case for it, so I shipped seven AI prototypes end to end. The model is never the hard part. Deciding what it should handle, and what stays with people, is.
Before product I studied Engineering Physics at IIT Delhi, which is where I picked up the habit of taking messy problems apart before solving them. Away from work I enjoy dancing and picking up new dance forms, collect logic puzzles, and protect a long Sudoku streak.
Education
Carnegie Mellon UniversityMaster of Integrated Innovation for Products & Services2025 – 2026 · Graduated May 2026
Indian Institute of Technology, DelhiBachelor of Technology, Engineering Physics2013 – 2017
Certifications
Google AI Professional Certificate↗Google · Jun 2026McKinsey.org Forward Program↗McKinsey.org · Jun 2026Vibe Coding: AI-Powered App Development for Product Managers↗University of Maryland · Jun 2026AI for Product Management↗Pendo.io · Jul 2024Introduction to Finance, Accounting, Modeling and Valuation↗Udemy · May 2020The Business Intelligence Analyst Course↗Udemy · Apr 2020
02 · Experience
Where I’ve worked.
Six years across fintech, e-commerce, logistics, and healthcare, mostly in India’s startup ecosystem.
Jun 2022 – Jul 2024PaytmManager, Product · Fintechcase study →Designed a daily-streak retention program that drove 5% DAU growth and a 20% transaction volume lift on tested cohorts. Won full authority over the shared app homepage and delivered a 40% CTR lift. Owned the full user lifecycle with P&L responsibility.Product Strategy · A/B Testing · Gamification · User LifecycleMay 2019 – May 2022ShopXGrowth Manager · B2B Commercecase study →Built the retailer channel from zero to 30K+ retailers on a B2B marketplace. Field research across 8 towns uncovered the messaging problem blocking adoption. Designed onboarding, stood up retention analytics, and improved on-time delivery from 85% to 94%.Growth Strategy · SQL · Python · TableauApr 2018 – Apr 2019ShipprProject Manager · Logisticscase study →Ran last-mile ops for Amazon and Flipkart, 370K+ shipments a month. Found damaged and misrouted shipments were the top cost leak, redesigned the warehouse, and trained ground staff. Losses dropped 30% and revenue grew 18%.Process Design · SOP Creation · Operations · Financial ReportingSep 2017 – Jan 2018Healers at HomeGrowth Manager · Healthcarecase study →First job. Hired 24 nurses, managed a team of 6, and built the sales scripts. Reworked channel messaging using competitor benchmarks. Bookings grew 30% and revenue 15% in 4 months.Sales Leadership · Process Automation · Growth Strategy · Customer Acquisition
03 · AI projects
What I’ve been building.
Seven builds from the past year, a mix of shipped products and prototypes, end to end. All code is public on GitHub.
AI Ticket Intake
An intake layer for MSP helpdesks. The AI drafts clean, structured tickets from vague requests; a technician approves before anything acts.
code ↗Collaborative Docs
A lightweight Google Docs, built for an AI-delivery take-home: rich text, sharing with revocable access, autosave. Deployed and usable.
live demo ↗ · code ↗Brand Builder
Visualizes a product across billboards, newspapers, and social posts while holding brand consistency, on Gemini's image model.
live demo ↗ · code ↗Holiday Hook
Finds obscure cultural holidays and micro-days that fit a brand's niche, then drafts daily content angles around them.
live demo ↗ · code ↗Data Dashboard
An interactive dashboard with analytical charts, filters, correlation heatmaps, and transaction editing.
live demo ↗ · code ↗Health Tracker
A personal iOS tracker for a PCOS-focused routine: supplements, exercise, trends, streaks. All data stays on the device.
code ↗04 · Decision log
What I chose, and why.
Trade-offs from six years of growth work and a year of building with AI, with the reasoning written down.
ProjectThe callChose / over
Paytm Homepage RevampAsk for full control of the homepage instead of optimizing a shared page.full authority with accountability / incremental optimization›
Why this call
Four teams shared the app homepage, each running their own banners, promotions, and A/B tests at once. My team was accountable for homepage visitors and transactions but had no control over what appeared on the page. The incremental path was safer, but optimizing a page four teams are editing is like mopping a floor four people are muddying at the same time. The problem was structural, not tactical. So I brought leadership scroll-depth data showing the page wasn't working, a forecast with specific numbers, and an offer: give us the page, hold us to these results, and take it back if we miss.
What it bought
A 40% click-through lift on tested cohorts, and a homepage that went from four teams competing to a single daily theme, a consistent layout, and a refresh cadence that gave users a reason to come back. The redesign was straightforward. Getting the authority was the hard part.
AI Ticket IntakeThe AI drafts the ticket. A technician approves it.human-in-the-loop / auto-close›
Why this call
A wrong auto-close at the front door of a helpdesk costs more than the minutes it saves, and after a few of them the technicians stop trusting the queue. So the AI drafts the ticket and suggests a fix, and a person confirms before anything happens. The model also has no data of its own; every suggestion comes from what the MSP already knows.
What it bought
Clean, reviewable tickets on a real 100-ticket export where 94 had landed in a single catch-all queue.
ShopX D2R LaunchShip two separate apps to distributors instead of waiting for a unified portal.two apps now, merge later / one polished app later›
Why this call
Distributors needed two different flows, one for placing orders to brands and another for receiving orders from retailers. Building both into a single interface would have pushed the launch by months, and the questions we had about retailer behavior, activation thresholds, and ordering patterns could only be answered by live users, not planning documents. The trade-off was real: two apps meant a worse distributor experience, some complaints, and an ops team supporting two systems for a while. But every week spent building the perfect portal was a week of assumptions instead of evidence.
What it bought
A working retailer channel in two sprints instead of four-plus months, and real usage data that shaped the unified portal, which shipped two months later and was better for it. The channel grew to 30K+ retailers.
The TiebreakerForce the model into a framework, not a free-form opinion.SWOT / comparison table / open answer›
Why this call
Nobody trusts a chatbot that just says "pick X." Putting the answer in a comparison table or a SWOT shows the reasoning, and the person still makes the call. The tool is there to break a tie, not to decide for you.
What it bought
Recommendations someone can check line by line instead of taking on faith.
Paytm GamificationKill spin-the-wheel after five failed variants. Move the budget to the daily streak.kill the format / a sixth version›
Why this call
Five versions of spin-the-wheel, with different rewards, triggers, and visuals, and none of them moved retention. Users enjoyed spinning. They smiled. But nobody came back the next day because of it. The daily streak created a behavior loop: come back tomorrow to protect your streak. The signal was already clear after version two; I held on for three more because the format felt fun and I kept thinking a different reward structure would fix it. It wouldn't have. The problem was the mechanic, not the prize. Now I set a kill threshold before the first variant launches: two misses on a format and the format is dead.
What it bought
Freed up budget and team time for the daily streak, which drove 5% DAU growth and a 20% transaction volume lift on tested cohorts. It also changed how I run experiments: decide in advance what "not working" looks like, so conviction can't overrule data.
Collaborative DocsOwn the auth and data layer instead of leaning on a managed default.Postgres + self-managed auth / Supabase Auth›
Why this call
Sharing was the core of the assignment, so I wanted the permission model in my own hands: owners grant and revoke access, passwords are scrypt-hashed, and nothing important hides behind a managed auth service.
What it bought
A live, deployed editor with autosave and real document sharing, built on an interview deadline.
Health TrackerHealth data never leaves the phone.on-device storage / cloud sync›
Why this call
It's one person's health data, so syncing it to a server adds risk and nothing else. SwiftData keeps everything local, and there's no account, no login, nothing to leak. The cost is no backup across devices, which felt like the right trade for a personal tracker.
What it bought
A tracker I actually use daily without thinking about where the data sits.
Holiday HookSkip the big holidays. Find the small ones.obscure micro-days that fit the niche / the major holiday calendar›
Why this call
Every brand posts on the big days, so those posts disappear in the feed. The days nobody else knows about are the ones a niche brand can own. The tool's job is to find them and suggest an angle, not to write the calendar for you.
What it bought
Content ideas a small brand wouldn't have found by hand.
05 · CMU projects
Product work from the master’s.
Complete semester projects at Carnegie Mellon, end to end: research, prototyping, and strategy in teams. Case studies on their own pages.
Capstone · Team Opus Mach @ CMU→
StepRight.Redesigning the pickup tailgate for every body that uses it.
Our capstone with Opus Mach, a precision engineering company: the tailgate redesigned as an access system, with a full-width step and grab handles for true three-point contact. One button, under five seconds.
06 · Notes
Notes from the work.
Short pieces from the work: what six years in growth roles taught me, and what a year of building with AI changed.
The career arc
From 6 People to 350 Million Users in Six Years
My first job lasted four months. My most recent one served a user base larger than the population of the United States. What each role handed to the next, and the three-step pattern that stayed the same the whole way.
Read · 6 min →From PaytmTwo Problems, Seven Experiments, and One Organizational Fight
Two transactions turned out to be the retention inflection point, one game format failed all five versions before I learned to set kill thresholds, and the homepage turned out to be an ownership problem, not a design problem.
Read · 6 min →From ShopXZero Retailers, Wrong Pitch, Eight Towns
Building the retailer channel from zero to 30K+: the two-app compromise, an onboarding system with a human in it on purpose, and the field research across eight towns that exposed the pitch as the real adoption blocker.
Read · 6 min →From Shippr370K Shipments and a Problem Nobody Could See
The company knew it was losing money but not where. Weekly loss tracking made the leak visible, and getting ground staff to see damage as money did the rest. Losses fell 30% and revenue grew 18% in eight months.
Read · 5 min →From a year of prototypesSeven small apps later
A year of building my own prototypes changed how I scope product work, and how I argue for it. Some of what I expected held up. A lot of it didn't, and that's the useful part.
Read · 5 min →From AI Ticket IntakeAI with no data of its own
Building the ticket intake changed the first question I ask about any AI feature: what data is it allowed to touch, and who owns that data? A note on why that question did more for trust than any accuracy number.
Read · 4 min →07 · Contact
I’m looking for my next product role. Say hello.
Recruiter, founder, or just curious about one of the projects: send a note and it lands in my inbox. I read everything.