I build web and mobile apps that work in the real world. Banks, delivery fleets, healthcare providers, they all run on software I shipped. 3 years, 10+ production apps, and a knack for making slow things fast.
Selected Projects
A social listening and customer engagement platform running in production at UBL and HBL. Microservice backend (Node.js + Python, Redis Pub/Sub, Docker Compose), integrated Meta Business API. Reduced message processing from 1.4 minutes to 0.4 seconds.
Won 1st at COMSATS Industrial Expo 2024 and 2nd at Securiti.ai AI Challenge. Flutter app with custom CNN models, Neo4j graph-based symptom analysis, and RAG treatment recommendations. Validated with ophthalmologists at WATIM Hospital.
Complete website and admin portal for SANHA Halaal Associates Pakistan. 50K+ monthly users. Cut page load from 4.2s to 2.3s through image optimization and API batching. JWT auth with role-based dashboards.
An analytics and workflow management platform used by Faysal Bank and Bank Alfalah. Helps banks understand customer behavior through data visualization and build dynamic operational workflows.
Offline-first water delivery platform built with Flutter and Node.js. Riders work without internet and sync when reconnected. Cut average fulfillment time from 25 to 15 minutes.
A social app for anime fans built with Flutter and Node.js. News feeds, discussion forums, social feed with posts, comments, and likes. Complete client and backend.
An invite-only social app for the vanlife community. Solves safety on the road, loneliness between stops, and decision fatigue. Built with Flutter, Node.js, and WebSockets for real-time messaging.
Awards & Recognition
Tools & Technologies
Work Experience
Blog
How I cut message processing from 1.4 minutes to 0.4 seconds
Optimizing a real-time NLP pipeline for Pakistan's largest banks using Redis, Bull queues, and Docker microservices.
Building offline-first apps: lessons from a water delivery fleet
Service workers, local caching, conflict resolution, and why it cut fulfillment time from 25 to 15 minutes.
What I learned building an AI eye disease detector with real doctors
CNN models, Neo4j graph search, RAG pipelines, and validating with ophthalmologists at WATIM Hospital.
Building AI agents that actually help: MedLens and what I'm learning
Real-time first aid AI using Google ADK, Gemini Live, and Vertex AI RAG. What works and what doesn't.
Connect
Feel free to contact me at awaisjarral37@gmail.com