Full-stack & AI engineer turning raw ideas into fast, resilient software — from the first commit to the first 25,000 users.
Founding engineerCherry Dating
25k+ usersacross shipped products
800+ problemssolved with intent
Selected workSelected experience
I like the part after the prototype: real users, real constraints, and systems that still have to feel fast on a bad connection.
Jan 2026 – Present
Architected a scalable mobile platform serving 10k+ monthly active users, sustaining sub-200ms response times handling 60k+ daily geospatial matching requests. Modeled an ELO-based recommendation engine increasing matches per user by 35%, and optimized AWS infrastructure caching to reduce monthly costs by 60%.
Oct – Dec 2025
Developed a scalable real-time chat system with WebSocket event-driven queuing to support 5,000+ concurrent users in production. Architected 20+ RESTful APIs optimizing response times by 30%, and refactored React state management to improve Core Web Vitals, decreasing page load times by 25%.
Mar – Aug 2025
Selangor, Malaysia
Shipped an enterprise full-stack ERP school management system securing 15,000+ records with strict RBAC. Optimized AWS RDS (PostgreSQL) through query indexing and connection pooling to cut API latency by 35%, and automated CI/CD pipelines via GitHub Actions and AWS Lambda.
Beyond the commit
Aarzoo Cultural Club
Entrepreneurship Cell
Evidence over adjectives
Consistently ranked in top percentile across global contests
Across LeetCode, CodeChef, and Codeforces
Scaled Cherry from 0 to 10,000+ users rapidly
1st Runner-Up in university-wide innovation hackathon
From design to deployment, shipping clean code is about keeping momentum. Hover to accelerate the build pipeline.
Working toolkit
Selected work
Proximity-based discovery app enabling real-time user matching within a selected radius for 1k+ daily location queries. Optimized spatial indexing in PostGIS using batching to achieve 48% faster match performance.
Engineered a search engine with Puppeteer and FlexSearch, cutting latency from 2.5s to <5ms for 50k+ documents. Integrated LLaMA 3.1 via Vercel AI SDK for real-time synthesized answers, boosting relevance by 40%.
Optimized Next.js prefetching to lift application load performance by 40%. Integrated Gemini API for automated, high-accuracy expense classification and budgeting across 1000+ financial transactions.
Available for the right team
I'm looking for product-minded engineering work with real ownership, demanding constraints, and people worth learning from.