I build full-stack software: web platforms, native mobile apps, cloud infrastructure. Same level of care whether it's the first line of code or what's running in production.
Live on the App Store: a production iOS app built in SwiftUI, with a custom Core ML / Create ML model doing real-time, on-device automotive recognition. Backend runs on Firebase, covering auth, data, and an admin panel for faculty service coordinators.
View on App Store →A multi-tier AWS setup end to end: EC2 for the web tier, S3 for static assets, RDS as the managed database layer, Auto Scaling Groups for elastic demand, and CloudWatch keeping an eye on all of it. Provisioned entirely as code through CloudFormation.
View on GitHub →A full-stack app built around object-oriented PHP: class-based architecture, clean MVC separation, and a fully normalized MySQL schema across four-plus tables with enforced foreign keys and real transactional integrity. Developed locally on a MAMP Apache/MySQL stack.
View on GitHub →A Swift iOS app built around Firebase Firestore as the backend: real-time CRUD, live sync across devices, and Firebase Auth handling users. A practical case study in cloud data architecture that holds up for production mobile systems.
View on GitHub →Designed, built, and published this one solo: a live iOS reference app pairing Urdu and English scripture side by side, with a custom layout that handles simultaneous right-to-left and left-to-right typography cleanly. It's fully offline, runs on a local on-device database, and ships with zero analytics and zero third-party tracking.
View on App Store →Sheet-resistance mapping for thin-film semiconductor samples, built with two teammates handling hardware and motor control. A motorized FUYU linear stage carries the sample under a fixed Cascade Microtech four-point probe, stepping 1mm at a time while an ADS1115 ADC on a Raspberry Pi reads voltage and computes resistance on the fly. I wrote the Python web server, the live charting UI, CSV export, and a demo mode that runs the full interface with no hardware attached.
View on GitHub →An aerial roof-inspection platform for iPhone, iPad and Mac, built around a computer-vision model that runs entirely on-device, keeping client imagery private by design. Inspectors import drone stills or flight video, review each finding, and export a branded, client-ready PDF report. Behind the interface sits a disciplined ML workflow: curated and license-audited datasets, controlled experiments, and quantitative evaluation on held-out imagery before any model reaches the app.
Native iOS and web development, and independent evaluation of AI and computer-vision models, delivered as clear written reports.