14 years engineering production systems.
8+ years specialising in Angular at scale.
Currently consulting from Bengaluru.
My work focuses on micro-frontend architecture, performance engineering, and the system-level decisions that make sustained, multi-team product development possible. I’ve shipped systems used by tens of thousands of pharmacies, decomposed a four-year monolith without freezing the business, and built a clinical platform end-to-end on my own.
Previously at Halodoc, Indonesia’s largest health-tech platform, where I led the micro-frontend migration of a four-year Angular monolith, owned production infrastructure for the public website, and mentored engineers from junior to mid-level.
Currently consulting independently from Bengaluru — providing architecture and technical advisory for engineering teams, and taking on full-product engineering from discovery to deployment when the fit is right. Angular at scale is my specialty, though the work isn’t framework-specific; and I deliver AI-native, so senior judgment isn’t bottlenecked behind hand-cranked implementation.
I work with scaling startups — typically Series A/B — whose Angular codebase has started to slow them down. Smaller teams usually bring me in to lead a migration end-to-end; larger ones more often start with a fixed-scope audit or a fractional advisory engagement.
A structured review of your frontend codebase — architecture, build and release bottlenecks, performance budgets, and the highest-leverage places to intervene — ending in a prioritised roadmap you can act on, with me or with your own team. It’s the lowest-risk way to start, and the best way for both of us to learn whether the bigger work is worth doing.
Leading or embedding with your team to actually do the migration — incrementally, validated in production, using a Strangler Fig approach so the business keeps shipping throughout. I run an AI-native delivery workflow, so the mechanical refactoring moves fast while the architectural calls stay deliberate. This is the work behind the build-time and velocity numbers in my case studies.
Senior engineering judgment for teams already in motion — architecture decisions, performance audits, technical code review, and direction on the hard calls. On tap, without a full-time hire.
I also take on full products end-to-end — discovery to deployment, across the stack, delivered AI-native. Most recently a clinical platform now serving 13,000+ patients.
Four years of growth had turned the codebase into the bottleneck.
Halodoc’s internal operations console — used across pharmacy delivery, tele-consultation, insurance, and labs — had grown over four years into an Angular monolith. Build times of 20–30 minutes, slow release cycles, and multiple teams blocking each other on a single codebase.
I evaluated four micro-frontend approaches through structured analysis and proof-of-concept, then chose Reverse Composition based on the application’s specific constraints: internal use, no SEO or SSR requirements, and Angular framework limitations around shared runtimes. We migrated incrementally using the Strangler Fig pattern — starting with low-risk modules, validating in production, iterating, then expanding.
Result: Build times dropped from 20–30 minutes to 5–8 minutes. Twelve micro-frontends in production. Independent team velocity restored. Zero feature freezes during migration.
Read the architectural write-up on Halodoc’s engineering blog →Three teams, one tightly-coupled module, a constant coordination tax.
A shared payment library sat at the centre of three independent product teams — and because the coupling ran deep, every change meant cross-team coordination before anyone could ship. I diagnosed the systemic coupling and led a cross-functional refactor that introduced clean separation of concerns, so each team could evolve its own payment flows without waiting on the others.
I also integrated a second payment provider, Xendit, alongside the existing Midtrans behind a facade abstraction — cards with a 3DS/OTP iframe flow, the LinkAja wallet via redirect, and virtual accounts via polling — designed so the consuming services never needed to know which vendor sat behind the interface.
Result: Onboarding a new payment vendor went from a coordinated multi-team change to zero changes in the three consumer services. Each team could ship payment work independently.
Building Halodoc’s pharmacy supply network — and removing a manual sales bottleneck. Led a team of four.
v1.0 — Architected and led the build of Medisend from scratch for 14,000+ Indonesian pharmacies. The central decision was a custom 3-layer state architecture (presentation → CartFacade → CartState) that kept the admin, seller, and buyer journeys cleanly separated as the platform scaled.
v2.0 — Led the merger of Halodoc’s existing Portal Apotek into Medisend, eliminating a manual activation bottleneck that previously required sales teams to make multiple in-person visits per pharmacy. Built self-service registration across four pharmacy types with role-based homepages.
Result: Sales teams freed from manual activation work. Platform serves the Indonesian pharmacy network end-to-end across both buyer and seller journeys.
End-to-end ownership of a production system — architecture to deployment, solo.
An Ayurvedic clinic in Gokarna with 13,000+ historical patient records had no digital workflow — clinical notes, prescriptions, and patient history all on paper. I designed and built a complete clinic management system from architecture through deployment, owning every layer of the stack myself.
The architecture was the point: an Nx monorepo sharing TypeScript contracts between an Angular front end and a NestJS/PostgreSQL back end, multi-role access for staff, doctor, and admin, and the full clinical workflow — including the Ayurvedic Nadi Pariksha consultation — through to bilingual English/Kannada PDF generation.
Result: In active clinical use, deployed on AWS LightSail behind a NetBird WireGuard VPN mesh. All 13,000+ historical patient records digitised and queryable.