AI-focused Computer Science student at the University of Malaya, building practical AI products, data tools, full-stack systems, and backend/data infrastructure.

• AI product engineering
• Machine learning and data analytics
• Full-stack web applications
• Backend, database, and data integrity
• Practical automation and tooling

Personal identity, document vault, proof-pack sharing, and AI authenticity checker MVP.

• Tech stack: Next.js, TypeScript, Tailwind CSS, Supabase.
• What I built: A full-stack MVP for managing personal identity materials, organizing documents, preparing shareable proof packs, and exploring AI-assisted authenticity checks.
• Why it matters technically: Shows product thinking around identity, trust, document workflows, privacy-aware data handling, and AI-assisted verification.

AI-assisted receipt extraction app that converts receipt images into editable structured form data with human review.

• Tech stack: Next.js App Router, TypeScript, Tailwind CSS, Gemini Vision API.
• What I built: Receipt image upload, server-side AI extraction, structured field parsing, editable review form, raw model-output preview, and local browser persistence.
• Why it matters technically: Demonstrates a practical AI workflow with server-side API boundaries, defensive model-output handling, and human-in-the-loop review.

Python toolkit for cleaning, searching, auditing, and analyzing makerspace inventory records.

• Tech stack: Python, CSV processing, CLI tooling, unit tests.
• What I built: CLI search, issue detection, duplicate detection, inventory summaries, visualization support, and sanitized sample-data workflows.
• Why it matters technically: Shows data cleaning, practical automation, privacy-aware dataset handling, and testable Python engineering.

Regression-based machine learning workflow for hourly KTM Komuter ridership prediction.

• Tech stack: Python, Jupyter Notebook, pandas, scikit-learn, XGBoost.
• What I built: A machine learning workflow for exploring temporal and station-based ridership patterns using regression models.
• Why it matters technically: Demonstrates applied ML thinking, feature engineering, model comparison, evaluation discipline, and honest limitation reporting.

Mobile-first membership portal prototype using React, Tailwind CSS, Vite, and Supabase.

• Tech stack: React, TypeScript, Vite, Tailwind CSS, Supabase.
• What I built: A membership portal prototype covering authentication, registration, profile management, and mobile-first user flows.
• Why it matters technically: Shows full-stack product development, auth/data modeling, responsive UI design, and Supabase-backed application structure.

Android/Flutter water quality monitoring client with sensor ingestion, analytics, GPS reporting, and backend integration.

• Tech stack: Android, Java, Flutter, Dart, Bluetooth, JWT-secured backend integration.
• What I built: Mobile monitoring flows for water-quality readings, historical analytics, GPS-based pollution reporting, and backend-connected app logic.
• Why it matters technically: Demonstrates mobile development, IoT-style data ingestion, backend integration, and real-world environmental monitoring workflows.

Languages: Python, TypeScript, Java, SQL / PL/SQL, Dart

Frontend: React, Next.js, Vite, Tailwind CSS

Backend & Database: Supabase, PostgreSQL basics, Oracle SQL / PLSQL, API route handlers

AI / ML / Data: Gemini Vision API, machine learning workflows, regression models, CSV/data cleaning, structured extraction

Tools: Git, GitHub, Vercel, Docker basics, VS Code, CLI workflows

• Testing and CI for portfolio projects
• Reproducible ML workflows
• Secure environment configuration
• Architecture documentation
• Production-style project presentation