My first job lasted 4 months. My most recent one handled a user base larger than the population of the United States. The path between them wasn't planned. But looking back, each role set up the next one in ways I didn't recognize at the time.
Healers at Home: learning to build the machine
September 2017. Three months out of IIT Delhi with an Engineering Physics degree and no product experience. I joined an at-home healthcare startup as Growth Manager and immediately got handed two problems: there's no sales team, and there's no sales process.
I hired 24 home nurses. Managed a team of 6. Wrote the sales scripts myself because nobody else had. Trained the team on objection handling. Pulled data across 11+ marketing channels, found the messaging wasn't working, compared against competitors, and reworked the copy on the three channels that mattered.
Monthly bookings grew 30%. Revenue grew 15%. In 4 months.
The role was short. The lessons lasted. I learned that growth starts with a functioning machine, not a strategy deck. You need people, a process, the right message, and a way to track what's working. Everything else is noise. I also learned I liked building things more than running things. Four months of building a sales engine from scratch was more interesting to me than the prospect of optimizing it for the next year.
Shippr: learning to see what's broken
I moved to a logistics startup doing last-mile delivery for Amazon and Flipkart. The scale jumped immediately. 370K+ shipments per month. 80+ delivery executives. Full P&L ownership.
Shippr taught me something Healers didn't: how to find the real problem. At Healers, the problem was obvious. No team, no process, build both. At Shippr, the problem was hidden. The company was losing money and nobody knew exactly where.
I built a weekly tracking system. Categorized every loss by type and route. The data pointed to damaged and misrouted shipments. Not the fleet. Not the fuel. Packages getting broken in the warehouse and loaded onto wrong trucks.
The fix required changing both systems and behavior. I could redesign the warehouse in a week. Getting ground staff to treat every damaged shipment as direct money loss took months. That gap between "fixing the process" and "changing how people think about the process" became a core part of how I approach problems.
I left after a year. Not because the role was bad but because I wanted to work somewhere the product itself was digital, not just the logistics layer around it.
ShopX: learning to build from zero
ShopX was a three-year role and the most formative one. I built an entirely new channel on a B2B marketplace. Zero retailers on the platform when I started. 30K+ when I left.
The skills from Shippr carried over directly. At Shippr I learned to make invisible problems visible through data. At ShopX I used the same instinct to build an analytics layer where none existed. Adoption rate, reorder rate, order frequency, drop-off, NPS. All tracked daily, all built from scratch.
The skills from Healers carried over too, in ways I didn't expect. At Healers I wrote sales scripts and trained a team on how to pitch. At ShopX I wrote onboarding playbooks and trained field reps on how to pitch retailers. Same muscle, bigger stage.
But ShopX added something new: user research. I went to 8 towns and sat in shops. Talked to retailers. Discovered that the company's messaging was completely wrong for the audience. That discovery changed my approach to every product problem after. Dashboards tell you what's happening. Users tell you why.
ShopX also taught me how to ship imperfect things on purpose. The two-app compromise for distributors. The manual onboarding system that wouldn't scale but worked for users who needed hand-holding. The decision to launch in 2 sprints instead of waiting for the perfect version. Each of these was a deliberate trade-off between speed and polish, and in every case speed won.
Paytm: learning to operate at scale
Paytm was the scale jump. 350M+ users. A competitive market against PhonePe and Google Pay. Real A/B testing infrastructure. Monthly P&L for engagement budgets.
I started in marketing, moved to product in 6 months, and owned the full user lifecycle within 4 more months. That progression speed wasn't accidental. Every role before this had trained me to walk into chaos and start building structure. Marketing to product in 6 months felt natural because I'd already done the equivalent jump at every previous company.
Paytm added the experiment layer. At Healers I fixed messaging by studying competitors. At ShopX I fixed messaging through field research. At Paytm I fixed engagement through structured A/B testing. Seven experiments across two game formats. Clear kill thresholds. Test vs. control. Cohort-scoped results. The instinct was the same, find what's broken and fix it, but the methodology became rigorous.
Paytm also taught me something about organizational problems. The homepage revamp wasn't a design challenge. It was a coordination failure disguised as a design challenge. Four teams sharing one page, each optimizing in their own direction. The fix was getting authority, not changing the layout. I wouldn't have recognized that pattern without Shippr, where I learned that sometimes the system itself is the bottleneck, not the people operating it.
The thread
Every role had the same three-step pattern. Walk in. Find the real problem (usually not the obvious one). Build the system to fix it and measure the result.
At Healers, the problem was: no machine exists. Build one. At Shippr, the problem was: the machine is leaking money but nobody can see where. Make it visible. At ShopX, the problem was: the machine works but the message is wrong for the user. Go find out why. At Paytm, the problem was: multiple machines are competing on the same surface. Get ownership and unify.
The scope grew. 6 people, then 80+ delivery executives, then 30K+ retailers, then 350M+ users. The tools improved. Excel macros, then dashboards, then SQL, then A/B testing frameworks. But the core approach stayed the same: don't trust the first diagnosis, build visibility before building solutions, and measure honestly.
What's next
I'm finishing my Master's at Carnegie Mellon and relocating to the Bay Area. The past year I've been building with AI tools, not because I want to become an AI specialist, but because AI is the next layer of leverage for the same growth problems I've been solving for six years. Personalized reward mechanics instead of manual cohorts. Automated experiment analysis instead of weekly data pulls. Conversational onboarding instead of human-on-every-call.
The instincts don't change. The tools get better.
Several of these roles have their own detailed note: 370K Shipments and a Problem Nobody Could See covers Shippr's operations story. Zero Retailers, Wrong Pitch, Eight Towns covers the ShopX D2R build. Two Problems, Seven Experiments, and One Organizational Fight covers Paytm's gamification and homepage work.