Most devs make it work. I make it feel right.
I came to engineering through the back door — premedical student, self-taught across five domains, then thrown into production systems alone. That path is why I build differently.
See the work ↓- Location Pakistan (Remote Ready)
- Experience Full-Stack / Mobile / AI
- Primary Stack React Native · Node · AWS
- AI Systems LangGraph · Context Eng.
- Domain Healthcare / Clinical
Core Projects
01 / WorkClinical Workflow Assistant
Lead DeveloperA high-stakes clinical decision assistant built from scratch — solo. Coming from a premedical background, I understood the clinical workflows I was engineering for, not just the code. Shipped production-ready apps to both iOS and Android stores, managing the full release cycle including TestFlight, Play internal testing, rejection iterations, and live submission.
The Encounters feature had no brief and no designer. I researched from the user's perspective, wireframed it, and shipped it. The client loved it. Also built the web portal (Astro) — a read-only interface allowing clinicians to copy summaries and encounters directly into EMRs, using the same login credentials as the mobile app. Briefly led a contracted React Native developer — onboarding, code review, and quality assessment.
AI Receptionist Dashboard
Full-Stack EngineerBuilt the dashboard foundation and serverless backend infrastructure within a 30-day window. Also delivered the marketing site in Astro. Shipped a working product foundation for the client to continue building on.
School Management System
Full-Stack LeadA comprehensive school management system built as a core product. Architected complex fee logic, multi-role administrative workflows, and data-intensive modules across a Next.js frontend and Node.js backend — maintaining code quality and architectural consistency throughout.
Professional Thread
02 / Growth PathExplorer Phase
Premedical student. Drawn to HTML, CSS, and JS in 9th and 11th grade with no curriculum pushing me there — just curiosity about what was possible on a screen. Built nothing serious. Absorbed everything.
The CSE Pivot
Gap year, entrance prep, BSc Computer System Engineering. Formalized the intuition with structure. Self-taught React, completed Andrew Ng's ML Specialization, then interned as a backend engineer — Django, FastAPI, then Express. No classroom for this part.
Real Constraints
The school management system was the first project with a real design system and real administrative complexity. Grew from backend contributor to full-stack lead, building fee logic and multi-role workflows that had to survive daily production use by non-technical staff.
Rapid Execution
The AI receptionist dashboard was a one-month sprint — dashboard foundation and serverless backend, delivered and handed off. First time working on a client product with a live domain and real users waiting.
Going Alone
The co-developer on the clinical workflow assistant left two months in. The client wanted chat. I added WebSockets, migrated from serverless to Fargate to EC2, rebuilt the AI layer in LangGraph, and managed the full store submission cycle — TestFlight, Play internal testing, rejections, and eventual live release on both stores.
The Encounters Moment
No brief. No designer. Time was short. I researched from the user's perspective, came up with the wireframe, and implemented it. The client loved it. That is the sentence that defines the shift from developer to product engineer.
I was a premedical student before I was an engineer.
That background is not incidental — it's why I could build the clinical workflow assistant with clinical context, not just technical competence. I understand the workflows, the terminology, and the friction points that clinicians face. Combine that with 8+ months of active development on a live healthcare product, and the domain knowledge is real, not assumed.
Across every project, I have been the person who keeps a check on code quality — architecture decisions, review standards, and the call on what ships. In this AI age, that judgment matters more, not less. AI generates code faster than ever. Someone still needs to know what good looks like.
Every stack shift in my work was driven by what the product needed, not what I already knew.
Currently working on the clinical workflow assistant. Based in Pakistan, remote-ready.