Eight years scaling the device-lifecycle platform behind Apple, Samsung, HP and AT&T — then two products for safari operators, built end to end, alone. I lead teams of up to ten and still review the diff.
Monolith → independently deployable Node.js services, migrated one capability at a time with the old path live.
Ten years, three cities, one consistent bet: the interesting problems live where software meets the people who have to keep it running.
I started in 2016 on live video — detecting ad breaks in IP streams and swapping in personalised spots. Then eight years on one platform as it grew from a consumer app for device repair into infrastructure that Apple, Samsung, HP and AT&T route real warranty and claims volume through. Mumbai, then Toronto, then Bengaluru.
Managing never took me out of the code. I have never been able to set architectural direction I wasn't willing to read line by line — and the months I spent building Halisi alone, every layer from multi-tenant isolation to 2am production support, were partly to check that the instinct was still real. It was.
Every figure came out of one rebuild, or the delivery discipline built around it. Ask me what broke on the way — that part is more interesting.
Video player → consumer app → enterprise platform → my own products. Each step a larger blast radius than the last.
Two products for safari operators, designed and shipped end to end as sole engineer — a headless itinerary CMS and a multi-tenant booking and operations platform. Both live, with paying operators running daily business on them.
Both products in detail ↓ section 04
The device-lifecycle platform behind Apple, Samsung, HP and AT&T — protection plans, claims, repair, exchange and buyback, running across four regions.
Booking product, e-commerce site and internal CRM for doorstep device repair.
Auto-detection of TV ad breaks on live IP streams, swapping broadcast ads for personalised video.
Four enterprise partners, four regions, zero acceptable downtime, and a roadmap nobody would let us pause.
One codebase served Apple, Samsung, HP and AT&T across India, MEA, EU and the US. Every release was a coordinated event, and a bug in one partner's warranty flow could take down another partner's claims flow. The incident rate reflected exactly that.
Not the architecture — the team. Ten engineers across four squads, each with their own partner commitments, could just as easily have built four inconsistent services and landed on a distributed monolith. That is worse than what we started with.
Refused the big-bang rewrite. Drew service boundaries around business capability, standardised API contracts before anyone wrote a service, then strangled the monolith one capability at a time — old path live until the new one proved itself. Observability, distributed logging and Jenkins pipelines went in ahead of the migration, so a deploy became boring rather than an event.
Four squads came out of it shipping to one standard rather than four. That outcome matters more to me than the incident graph.
Halisi Safari Solutions is safari-operations infrastructure. I designed, built, shipped and support both products — architecture, data model, frontend, auth, billing logic, deploys and the 2am incidents.
A self-hosted headless CMS and marketing frontend for safari itineraries. Editors stack blocks into pages and multi-day trips, and the same content renders as a web page or a print-ready PDF.
The operations platform underneath. Each operator signs in on their own subdomain with their own branding, users and roles; the Halisi console runs on the apex domain and onboards new tenants.
Both repositories are private under the Halisi-safari-solutions organisation — happy to walk through the code in an interview.
Everything on this page, queryable. A few commands aren't on the list.
If you're rebuilding something in flight, standing up a platform team, or you need someone who will lead it and still write code — I'd like to hear about it.
Open to Engineering Manager and Senior / Staff Engineer roles. Remote, hybrid or relocation.