About
I've been building things since 2011. Not managing things. Not planning things. Actually building them — writing the code, debugging the hardware, shipping the feature, watching real users use it. Fifteen years in, that hasn't gotten old. If anything, it's gotten more interesting.
This is not a resume. It's a story of what I've actually built and why I'm not done yet.
The Journey
It started with Android. In 2011, when the ecosystem was still young and a decent mobile app felt like magic, I was writing apps and figuring out how to make things work on devices that had barely any RAM and a fragmented OS. That time taught me something I still rely on: when the platform is immature, your fundamentals have to be solid.
From mobile, I moved into fintech — and fintech changed everything. At Mahindra Comviva, I worked on Zoto, which became Nigeria's leading payments super-app. Millions of users sending money, paying bills, accessing financial services from their phones. The scale was real, the stakes were real, and the complexity of building in a regulated, multi-currency, high-transaction-volume environment taught me what engineering actually means when it has to work.
That same period produced DBXP — a digital banking platform that went live across 15+ countries. Not a prototype. Not a pilot. A platform in production, handling real money, in markets with different regulations, different currencies, and different user behaviors. Building for that kind of breadth forces you to design systems that are genuinely flexible rather than just theoretically clean.
Then came logistics. I joined Raaho, a logistics tech startup, and took on a different kind of challenge — leading cross-functional engineering teams, architecting a backend of 50+ microservices, and building in a domain where the physical and digital worlds collide in ways that keep you humble. Trucks move. Drivers go offline. GPS drifts. Systems that work perfectly in staging break in interesting ways when a driver is in a dead zone on a highway at 2am.
Along the way, I built a complete closed-loop NFC Tap & Pay prepaid payment system — alone. Hardware integration, backend API, transaction engine, merchant dashboard. The whole thing. That project taught me more about end-to-end systems thinking than anything else I've done. When you're the only person who can fix something, you learn every layer of it.
Then came AI agents. The most recent one is Zenith — the help and support chat inside a live-events ticketing app. Claude runs a bounded tool loop over a user's own bookings: it finds the ticket, explains the refund rule, drafts the support ticket. It can read what it needs and change nothing on its own; refunds and tickets wait for the user's tap. Around it I built the parts that make an agent safe to run and change: an ops dashboard, a replay lab that re-runs past chats against new code, and a cost ledger per feature. The full write-up is here.
Before Zenith there was Chhotu Bot — a Telegram assistant powered by OpenAI — which handles GitHub PR summaries, CleverTap analytics reports, and DigitalOcean cron monitoring for my team. What used to be 90 minutes of daily dashboard-checking is now three commands in a chat window. I built it because I was tired of being the glue between my own tools. I kept building it because it actually worked.
What Drives Me
I build because I genuinely cannot stop thinking about what's broken and how it could be better. Not in an abstract, whiteboard-session way — in a "why is this dashboard still open when I could automate this" way. The thing that gets me out of bed is the gap between how things work and how they could work, and the fact that I know how to close it.
I'm most alive when the problem is hard, the domain is real, and I have the freedom to go from idea to production without asking permission at every step.
What's Next
I'm currently a Principal Product Engineer at Raaho, building real systems that move real trucks across real roads. I'm also looking for my next role: principal or staff-level engineering at an AI-first company, where intelligent systems are the actual product, not a roadmap item. Remote or global, I'm open to both.
What I bring is 15 years across fintech, logistics and distributed systems, and recent hands-on work building AI agents that run in front of real customers. I'm most useful where the problem is hard and the domain is real.
Let's Connect
If you're hiring for that, or building something in that space, I'd like to talk.
LinkedIn or gaurav.arora90@gmail.com. My resume is here.