Founder. Chief of staff to a CEO. Product leader at Zomato, Amazon and Careem. Over ten years building for tens of millions of customers across India and the Middle East.
MBA projects with Pulsar Products (USA), Mary Kay (China) and Tata Tinplate (India).
Founder, early employee, chief of staff and product leader, across logistics, travel, food delivery, e-commerce, gaming, health and a super-app.
Pick the closest fit. The headline, the evidence and the order of the case studies change to match.
Roles that own a marketplace, partner platform or contributor network: onboarding and listing, payouts and incentives, trust and quality on both sides.
Roles that take AI from demo to daily use: choosing the problem, proving it against a baseline, rolling it out safely and measuring adoption.
Roles that sit with a client or a business unit and deliver change: AI transformation, forward-deployed product work, strategy that ends in something running.
Roles that own acquisition, activation, conversion and revenue: freemium to paid, pricing and fees, experiments that decide what ships.
Roles that own the consumer experience, loyalty and retention: why people come back, and what they're rewarded with when they do.
Roles in e-commerce, food delivery and quick commerce: selection, sellers, fulfilment speed and unit economics.
Roles in health tech, wearables and consumer apps: privacy and consent, daily habits, and partners in the care system.
Roles where the product runs on fleets, devices, networks or deliveries: operations at scale, safety, and rollout across many markets.
Roles in consumer fintech, payments and spend management: wallets, subscriptions, payouts and expenses that people actually enjoy using.
Head of, VP and executive roles: owning a P&L, setting the bar for a product org, and staying hands-on while doing it.
Three cases to start, closest to your role if you picked one. Open any for the problem, the decision, the mechanism and the result.
Careem had payments, discovery, a user funnel and advertising. It did not have fulfilment for the long tail of services people want. Tenants let local businesses plug into everything else.
Careem had payments, discovery, a user funnel and advertising, but no fulfilment for the long tail of services people want. Building every vertical in-house (car rental, laundry, home cleaning, pet grooming, gardening, e-bikes) would never pay back. Car rental ran through Swapp, a partnership live since 2022 that had become strained.
Bring the long tail onto Careem without funding new verticals, and turn the Swapp partnership around.
Support tickets in the last five days of every month were dominated by one request: tax invoices. The people asking were being reimbursed by employers. So the employer became the customer.
While I ran a six-month care-automation strategy, agent-only tickets rose from about 10% on a normal day to about 70% in the last five days of each month, in the UAE, KSA and Jordan alike. The cause was tax-invoice requests: Careem issued receipts, not tax invoices. Interviews with 15+ users found employers reimbursing 20 to 40 trips and orders a month; real-estate agents were the surprising segment.
The idea had been shelved before. Prove that companies would pay, then make them a paying side, while still delivering the care-automation goals.
Photograph a list, paste one forwarded on WhatsApp, or speak to an agent. Version one ran on human transcribers, so the model had a baseline to beat before it shipped.
Quick commerce trains people to top up, not stock up, and a weekly shop took about 40 clicks. Raising grocery order value was the one goal every vertical declined at the company hackathon I created.
My team took that goal: raise order value without buying it with discounts.
Food discovery leaked at every step, and office-goers kept ordering from the same handful of restaurants. So we built for habit instead of basket size: complete set meals from nearby favourites, checked out in one tap.
Food discovery leaked at every step: about 30% of users lost from search to cuisine, 50% from cuisine to restaurant and 40% from menu to checkout. Most office-goers, many of them expats who don't cook, ordered from the same 10 to 12 restaurants near where they live or work.
Build for repeat and habit instead of basket size, against a 40% three-day-repeat target set before launch, and bring two teams with order-value targets along.
A charter that started as three people and became the implementation partner for every vertical, now building on Claude Code.
Process knowledge sat across Confluence, Notion and Drive, restaurant onboarding took six days, and every Careem vertical needed design, engineering, PRD and data help to put AI to work.
Pitch and found an AI Enablement charter, starting as three people, and make AI part of how work gets done across the company.
As co-founder and CPO, I built a consent-based health-records layer on India's ABHA rails, the way payment apps sit on UPI. It only worked if five partner groups took part: the government, insurers, hospital chains, diagnostic labs and private clinics.
Patients in India carried paper files between hospitals that would not accept each other's reports, so tests were repeated and no doctor saw the full history.
As co-founder and CPO, build a consent-based health-records layer on India's ABHA rails, the way payment apps sit on UPI, and bring every side of care onto it.
As Chief of Staff to the CEO, I ran the pivot from messaging to gaming and the economy that paid for it.
Hike was pivoting from messaging to gaming and needed an economy that paid for the move.
As VP and Chief of Staff to the CEO, run the company internally, lead the pivot and own the game economy.
I visited micro-merchants in their stores. The new ones were lost, and many went to cyber cafes to use a web-first seller portal. So we rebuilt it mobile-first, from India for the world.
Micro-sellers were behind 80% of India volume, and many went to cyber cafes because the seller portal was web-first. Buyers could not tell real D2C brands from fakes, and fraudulent sellers reached customers.
As single-threaded owner of seller success, double onboarding success, measured on a seller's first 100 orders.
The north star was "shutdown", the time the app showed nothing deliverable, and the goal was zero. Paying more bought only marginal supply. Unlocking new supply changed the curve.
80 to 85% of peak orders were within about 3 km, but serviceability rules switched off long-distance restaurants at peak, and requiring a motorcycle filtered out supply. Milk and newspaper riders on cycles sat idle at lunch and dinner. Paying more bought only marginal supply.
Drive “shutdown”, the time the app showed nothing deliverable, to zero, at a cost per order that pays.
One payout structure per city could not pay a two-hour cyclist and a twelve-hour biker fairly. We rebuilt payouts in six phases, down to rate cards for cohorts as small as 10 riders.
One payout structure per city could not pay a two-hour cyclist and a twelve-hour biker fairly. Customers had to do the maths across bank, wallet, partner and Gold offers, and abuse ate into margin, all while Zomato grew from 5 to 500 cities.
Lead the fulfilment product pod (assignment, serviceability, payouts, fraud and cash collection) and own the unit economics of a 250-city P&L on a $20M budget.
Honeymoons to island destinations: high ticket, very high acquisition cost, and about 5% repeat over nearly five years. One purchase could not pay for the customer who made it.
Honeymoons to island destinations: high ticket, very high acquisition cost, and about 5% repeat over nearly five years. One purchase could not pay for the customer who made it.
As Product Leader, Customer Experience, own the experience across travellers, local travel agents and on-ground providers, and make travellers come back.
Small e-commerce sites would launch, sell through short-form content and close within three or four months, which was exactly the credit line logistics extended.
Small e-commerce sites would launch, sell through short-form content and close within three or four months, exactly the credit line logistics extended. And Indian addresses are free text, so one door arrives written many ways.
As one of Delhivery's first 30 employees, stop the credit leakage and make addresses routable.
The role, line by line, against the work above.
I've been the platform (Careem), the seller side (Amazon), the operator paying supply (Zomato) and the matchmaker between three sides (Travel Triangle). Partners and buyers are both people I've built for.
Prepaid B2B wallets, rider payouts, company spend policies, offer stacking and platform fees. A marketplace is only as trusted as its ledger.
I ship with Claude Code, publish skills, and built this site, my last fake-door pages and several products end to end myself. I would be a Marketplace partner's PM and its customer at once.
| The role asks for | Where I've done it |
|---|---|
| Own the marketplace end to end, onboarding and listing through purchase and billing | Tenants (vetting, commercials, listing on shared rails) and Careem for Business (admin workspace, funded wallet, checkout toggle) |
| Commercial workflows that meet finance, audit and compliance needs | DMiles prepaid B2B wallet, Zomato payouts built with the finance org, company spend policies at Careem |
| Commercial incentives built into the product | Cohort rate cards at Zomato, the referral loop at Travel Triangle, the Hike economy |
| Built with or shipped on LLMs | AI grocery list, Ask Careem, a skill published on Careem's internal skill marketplace |
| Go-to-market and partner deals | Swapp exclusivity and equity deal at Tenants; selling with and to enterprises at Careem, Amazon and Delhivery |
| Partner, app store or ecosystem experience | Platform side at Careem, seller side at Amazon, operator side at Zomato |
| Zero to one | Tenants, Careem for Business, the Hike pivot, Zealth |
| Define success metrics | "Shutdown at zero" at Zomato, trips per converted traveller at Travel Triangle, new accounts at Careem for Business |
What share of a partner's new enterprise revenue does the Marketplace source, and is that share growing?Measured before: Tenants partner GTV, ~5% to ~96%.
When committed spend goes to a partner, both finance teams need the same record of what was drawn, against which contract, and when. The ledger is part of the product.Built before: DMiles and the Careem for Business wallet.
How long from a partner submitting to an enterprise buying through procurement, and which step loses the most partners?Cut before: merchant onboarding, 6 days to 1.
A partner program with one set of terms will fit the median partner and nobody else. Which cohorts need their own terms?Built before: Zomato rate cards for cohorts as small as 10 riders.
These are questions, not claims about how the Marketplace works today. Facts about it come from the linked public sources.
Tell me about the role, or better, the problem everyone around you has written off. Based in Dubai, open to relocation.
Nothing here is needed to understand the work above. Open whatever you want to verify.
Searches the case studies on this page. Every answer names where it comes from.
Every item from my CVs and interviews in one place, 59 entries in all. The sections above tell the best stories; this is the complete list, and the Ask box searches all of it.
Scheduled orders on Careem Food, so customers can order ahead.
No number on record yet
No number on record yet
No number on record yet
No number on record yet
No number on record yet
No number on record yet
No number on record yet
Exchange programmes at Tongji University, China, and Case Western Reserve University, USA. Live projects with Pulsar Products (USA), Mary Kay (China) and Tata Tinplate (India).
Strategy and Marketing. President of India Scholarship. Founded Unnayan, which gave village students free tuition from student volunteers.
A number appears only where it is consistent across my own CVs and interviews. Where my own tellings differ, the item is listed without the figure.
ShippersTravellersRidersSellersPlayersPatientsPartners
Every move put me on a different side. The job never changed: make that side want to stay, and prove it with a number.
Tenants, Careem for Business, Careem Plus and the AI Enablement charter. Seven direct reports: five PMs and two APMs.
Health-records network on India's ABHA rails: one of the first five companies in the government sandbox, with government, insurer, hospital, lab and clinic partners.
Messaging to gaming pivot and the Rush economy.
Seller onboarding, Top Brands, the seller quality engine, founding team of India Ultrafast Fresh.
Fulfilment, rider payouts, promotions and fraud through growth from 5 to 500 cities.
Repeat and referral loop, weekend getaways, lead scoring.
Among the first 30 employees. DMiles and AddFix.
Top 5% of batch. Exchange at Tongji University, China, and Case Western Reserve University, USA. Live projects with Pulsar Products in Cleveland, USA (a subscription kit product), Mary Kay in Hangzhou, China and Tata Tinplate in India.
Airtel 3G and Airtel Money activation in villages around Jamshedpur; retailer analysis from 500 retailers.
Top 1% of class, President of India Scholarship. Founded Unnayan, which gave 7,000+ village students free tuition from student volunteers.
Two ways to see the same career: who I built for, and where. Pick any part to see the work behind it.
| Industry | Marketplace & platform | Growth & loyalty | Money & payments | AI & data | Operations |
|---|---|---|---|---|---|
| Super-appCareem | |||||
| Food deliveryZomato · Careem Food | |||||
| Grocery & quick commerceCareem · Amazon | |||||
| E-commerceAmazon | |||||
| TravelTravel Triangle | |||||
| GamingHike | |||||
| HealthZealth | |||||
| LogisticsDelhivery | |||||
| Mobility & home servicesCareem | |||||
| Enterprise & B2BCareem · Delhivery |
31 intersections with shipped work. Tap a dot to open it in the full record.
AI runs through all of it: the grocery list, AI Meals, Ask Careem, merchant onboarding and the agents behind Careem's operations.
I'm rarely handed a spec. I'm handed a problem area and a goal. I find what can be built, ship it with metrics, remove pain points along the way, and re-architect when the old design won't scale.
Grow Careem beyond the verticals it runs itself, and run all of it leaner with AI.
The brief was a direction, not a spec. Four pillars came out of it, and every product below ladders up to one of them.Partners plug inMore to buy in one appAdopters triple their spendPartners grow with Careem (~5% to ~96% of GTV)More partners want in↺
Drive “shutdown”, the time the app shows nothing deliverable, to zero, at a cost per order that pays.
One north star, four pillars. No single product could move it; the portfolio did.Short trips go to cyclistsMore riders at peakLess shutdown, more ordersCost per order ~₹120 to ~₹78Room to pay each cohort fairlyRiders stay↺
Pick any product to see what it did and how it moved the goal.
Each one came from a project that would have gone worse without it.
Minimum lovable, not minimum viable. A narrow scope, finished properly, beats a wide one nobody comes back to.
The grocery list did one thing well: 65% of adopters were back within two weeks.The interesting work is usually the goal other teams have already decided can't move.
At Careem's AI hackathon every vertical passed on grocery basket size. We took it.A working thing ends a debate faster than a better slide.
A prototype I built on Lovable settled the grocery list design; a fake-door page proved Careem for Business in two weeks.Discounts move a number this month and train customers forever.
Instead of a big-basket discount, country-of-origin tags made quality visible.I build it, run it until it works, and leave it with the team that owns it.
Ask Careem reached ~98% accuracy and regenerates its own documentation as features ship.Every partner product I've owned had a commercial motion attached, and I carried it until a sales team could.
Signed Tenants partners on agreed commercial terms across three countries, turned around a strained partnership with Swapp into an exclusivity deal with equity for Careem, and built lasting partnerships with Justlife and Washmen, and sold Careem for Business to its first companies myself.
Owned seller success for the micro-sellers behind 80% of India volume, and built Top Brands so D2C brands could sell on trust signals.
Moved B2B shipper accounts onto the DMiles prepaid model, and AddFix, our address engine, was later offered to Amazon as a service.
With Claude Code, Lovable and Replit, the distance between an idea and a working product is mine to close. Each one in the same four steps.
No usage numbers yet.
Built for my own family. No usage numbers yet.
Built for my own use. No usage numbers yet.
At university I founded Unnayan, which gave 7,000+ village students free tuition from student volunteers. It was my first marketplace: volunteers on one side, students on the other.
At university I co-hosted a quiz with Barry O'Brien for an audience of 100,000, was the only student on the committee for the 150 years of Tagore celebrations, and won state-level debating and business-plan competitions.
Brand design, and orchestrating several AI models and agents to work together. This site was built with Claude.
Four levers, four goals. It's a toy, but the trade-offs are the ones I worked through at Zomato.
You run food delivery in one city. Hit all four goals at once: 5,500+ orders a day, delivery in 30 minutes or less, riders earning ₹900+ a day so they stay, and no loss on each order.
You balanced it. Here's what happened for real.
At Zomato, most orders came from within about 3 km of the customer, so I built a cyclist fleet for short trips: 90K cyclists from zero, carrying 43% of orders at 50% lower cost, taking cost per order from about ₹120 to about ₹78. Rider pay moved to rate cards for cohorts as small as 10 riders. Across 250 cities the P&L went from -20% to +16% EBITDA.
Read the full caseA toy model. Its numbers are illustrative, not Zomato data. The results in the green box are real.
Saurav Patro