Interactive Terminal
Details
Skills & Technologies
Java Ecosystem
Database & ETL
Backend & APIs
Frontend & UI
DevOps & Tools
AI & Machine Learning
Work Experience
Kollect System Sdn Bhd
Java Developer, AI Engineer & Team Lead
Aug 2025 – Present | Kuala Lumpur
Murifa Technology and System Sdn. Bhd.
Full-Stack Analyst Programmer
Aug 2024 – Jul 2025 | Petaling Jaya, Selangor
Silicone Zone Networks
Web Developer
2022 – 2024 | Remote
Wash & Wear Laundry
Multimedia Specialist & IT Technician
2022 – 2024 | Remote
Tune Tech Voiture Sdn. Bhd.
Multimedia Specialist & IT Technician
2017 – 2019 | Remote
Education
Asia Pacific University of Technology & Innovation
Master of Science in Artificial Intelligence (Part-Time)
2025 - 2027 (Expected)
Currently enrolled, focusing on advanced AI, deep learning, and analytical problem-solving
Enhancing expertise in enterprise solutions and intelligent systems
Researching AI applications in financial and enterprise software
SEGi College Kuala Lumpur
Bachelor of Science (Honours) Computing - Upper Second Class Honours
2021 - 2024
Specialized in enterprise application development and system integration
Developed multiple projects using Java, Flutter, React, and Next.js
Strong focus on banking, inventory management, and financial software solutions
SEGi College Kuala Lumpur
Foundation in Information Technology
2020 - 2021
CGPA: 3.78, completed with academic distinction
Gained fundamental knowledge in IT principles, programming, and system design
Prepared for advanced studies in computing and software engineering
Sekolah Menengah Kebangsaan Hillcrest
Sijil Pelajaran Malaysia (SPM)
2019
Focused on science and mathematics curriculum
Active in school technology competitions and extracurricular activities
Projects
Kollect AI ETL — Enterprise Data Pipeline Platform
- $Built from scratch — production ETL platform using Java 21, Spring Boot 3.4, and clean architecture with adapter SPI pattern, supporting DB-to-DB and CSV-to-DB pipelines for ERP and Banking clients
- $Spring Boot backend with pluggable adapter SPI (CsvFileAdapter, JdbcDatabaseAdapter, RestApiAdapter) — SOLID principles, schema validation, type compatibility checks, and connection validation engine
- $AI auto-mapping via local Ollama LLM — analyzes source schemas and target database tables, suggests field mappings at 90%+ confidence with direct-load fan-out to 8+ target tables
- $SpecKit-driven development — feature specifications auto-generate implementation plans and task lists, with acceptance scenarios driving 25+ custom pipeline components
- $2-level orchestration system — Level 1: individual pipeline designer with 25+ node types, Level 2: workflow orchestrator composing pipelines into dependency-ordered batches
- $Real-time WebSocket (STOMP) execution monitoring with state machine (INIT → STAGING → TARGET → SUCCESS), per-node timing, row counts, and millisecond-precision audit trails
Architecture Highlights▼
• Spring Boot 3.4 backend with adapter SPI pattern: CsvFileAdapter, JdbcDatabaseAdapter, RestApiAdapter — pluggable source/destination system following clean architecture
• Java 21 with SOLID principles, OOP design patterns, and multi-module Maven project structure
• SvelteFlow (@xyflow/svelte 1.5) with 25+ custom node types: control flow (START, END, IF_ELSE, PARALLEL), sources (CSV, Database, API), transforms, destinations, and observability nodes
• 5-step Smart Flow Wizard: Source selection → Column analysis → Target discovery → AI mapping preview (confidence scores) → Confirmation and apply
• AI mapping cache (24h TTL) with multi-model support: Qwen3:8b for fast mapping, Qwen3:30b for deep analysis, name-similarity fallback when AI unavailable
• Connection validation engine: type compatibility checks, acyclic graph enforcement, port matching (RECORD_STREAM, FILE_STREAM, EXECUTION_CONTEXT)
• WebSocket (STOMP) real-time execution logs with state machine: INIT → FILE_PROCESSING → STAGING → TARGET → ARCHIVING → SUCCESS
• PostgreSQL schema design with migrations, query optimization, and comprehensive data validation pipelines
KollectGPT — Multi-Agent Text-to-SQL AI Platform
- $4-service microservices architecture — Spring Boot WebFlux gateway (JWT + AES-256-GCM encryption), FastAPI AI backend, SvelteKit 5 client + admin, Docker containerization
- $Enterprise AI platform serving Banking, ERP/POS, and Collections domains — natural language to SQL across 50+ tables with sub-second response times
- $Multi-agent ReAct architecture (LangGraph) with dynamic tool registry (30+ tools) — agents reason, call tools, observe results, and iterate until the answer is complete
- $5-layer caching strategy with Redis — Schema singleton, Query cache, Semantic cache, Redis sessions, FAISS vector index for sub-second responses across 500+ columns
- $Schema RAG pipeline with FAISS embeddings + FK graph traversal for automatic multi-table join detection and intelligent schema discovery
- $ML pipeline: XGBoost for payment prediction, scikit-learn for risk profiling, Prophet for time-series forecasting, spaCy for NLP context extraction
Architecture Highlights▼
• Spring Boot WebFlux gateway — JWT authentication, AES-256-GCM datasource encryption, REST API routing, rate limiting, audit logging
• Database-driven domain adapters — each domain (Banking, ERP/POS, Collections) has custom system prompts, tools, glossary, metrics, and followup patterns
• Zero-hardcoding tool registry loaded from database — tools filtered at runtime by domain type, schema validation, and feature flags
• Dual-path query engine with 4 migration modes (off → shadow → partial → full) — shadow mode comparison in production without user impact
• Conversation memory (Redis) with 10+ turn context preservation, entity tracking, pronoun resolution, and semantic follow-up query optimization
• Multi-provider LLM abstraction: OpenRouter, Ollama (air-gapped), OpenAI, Claude — with circuit breaker failover and automatic fallback
• Security: SQL injection prevention, read-only enforcement, Docker isolation, Prometheus metrics export
• Monitoring: shadow mode comparison dashboard, slow query detection, cache hit/miss tracking
AUTO INDUSTRIES FACTORY LLC Website
Next.jsTypeScriptTailwindFramer Motion
- $Designed and developed a professional corporate website for a UAE-based automotive fabrication company
- $Created an immersive hero section with animated logo and text content
- $Built an interactive services showcase with animated cards
- $Implemented a filterable projects gallery with smooth transitions
- $Designed a dynamic “About Us” section with company history
- $Added comprehensive SEO optimizations including metadata and structured data
Technical Highlights▼
• Server-side rendering with Next.js for optimized performance
• Type safety with TypeScript for better developer experience
• Custom animations using Framer Motion for engaging user experience
• Responsive layouts with Tailwind CSS and custom design system
• Accessible UI components with proper ARIA attributes
• Optimized images and assets for faster page loading
• Dark/light mode with system preference detection
QR Code Generator
Generate QR Codes with custom colors and size
Preview QR Codes before downloading
Download QR Codes in PNG, JPEG, or PDF
History tracking for up to 10 QR codes
Snake Game
Classic Snake Game with modern UI
Multiple theme options (Classic, Neon, Retro)
Dynamic difficulty scaling with score
Power-ups system with special abilities
2048 Game React
Classic 2048 puzzle game mechanics
Smooth animations with Framer Motion
Touch support for mobile devices
Score tracking with local storage
Resume
My Professional Resume
Get a complete overview of my skills, experience, and education
Feel free to download my resume for a comprehensive overview of my professional background.
