GitHub Repo
Reading List
My favourite reads on worldly stuff around tech and creative spaces.
About
I'm a software engineer with a background in AI/ML and computer vision. My work spans the full stack: end-to-end systems design, from model training to application development, with some hardware integration along the way.
The human side of engineering is something I genuinely enjoy: unpacking what users actually need and finding ways to make technical concepts land for all people.
Interests
Flexible monkey, but I'm particularly fascinated by these topics:
Computer vision systems
Setting up vision and software systems (motion capture, multi-camera, depth sensors) — 2 years specialisation in uni.
Edge AI deployment
Optimizing and running inference on local hardware for applications with strict latency, cost, and data privacy requirements.
AI performance engineering
Maximizing model efficiency by leveraging hardware acceleration and low-level software optimizations.
Certifications
Click a certification to verify.
Cloud
AWS Certified Solutions Architect – Associate (SAA-C03)
Mar 2026
HPC
Accelerated Computing in Modern CUDA C++
Feb 2026
AI/ML
Google Cloud Professional Machine Learning Engineer
Jan 2026
AI/ML
AWS Certified Machine Learning – Specialty (MLS-C01)
Dec 2025
AI/ML
Microsoft Azure AI Engineer Associate (AI-102)
Dec 2025
AI/ML
AWS Certified AI Practitioner (AIF-C01)
May 2025
Skills
Languages
Python
C++
C#
TypeScript
JavaScript
Java
SQL
Libraries
Semantic Kernel
PyTorch
NumPy
Pandas
OpenCV
Semantic KernelSvelte
.NET
Cloud
AWS
Azure
Tools
Nsight Systems
Git
Docker
Linux 🐧
Kubernetes
Nsight Systems
Others
Rhino
Arduino
RhinoCore Beliefs
- Architecting the Execution — AI handles the execution, but I own the architecture. My role as an engineer has evolved from writing syntax to designing constraints: defining explicit requirements, mapping domain logic, choosing the right stack, and enforcing architectural boundaries. By offloading the generation, I focus my energy where it matters most: bridging the gap between the product owner's vision and structural integrity through rapid, intentional iteration.
- Engineers Are Bridges — I see engineering as a bridge between people and technology. Building software is only part of the work; understanding human needs and translating them into technical solutions is where real impact happens.
- Hardware Grounds Software — Understanding the silicon beneath the abstraction keeps an engineer honest about what computation actually costs. Every layer leaks eventually; knowing what lies underneath makes better decisions possible.
- The Spirit of Open Source — I have a deep appreciation for the open-source community. I love studying open codebases to learn best practices.
Projects
AI Combat Manikin ⭐ Featured
Software Design | Computer Vision | Hardware Integration
Developed a modular AI-driven training system integrating a user interface with a Python backend with Intel RealSense depth cameras for real-time baton tracking and responsive manikin maneuvers.
Exploring Nanochat
LLMs | PyTorch | CLI
Explored Andrej Karpathy's nanochat, where one can train their own LLM for cheap. I trained and customised models at different depths on my own GPU.
SUTD Q&A Chatbot
HuggingFace | LLMs | RAG
Developed a Retrieval-Augmented Generation (RAG) chatbot prototype using Llama 3.2 to answer SUTD-related questions, with an LLM-as-a-judge evaluation system.
CrowdCare
IoT | K3s | Arduino
IoT system for real-time public health monitoring using HiveMQ and ESP32 modules. Leverages K3s on Raspberry Pi for scalable, automated container management.
Interlocking Tiles Defect Detection GUI
Computer Vision | Python | YOLOv8
A no-code interface leveraging YOLOv8 for efficient defect detection.
Road Sign Detection System
Computer Vision | PyTorch | OpenCV
A robust solution for recognising Singapore road signs for autonomous vehicles.
Pixel Run
FPGA | Lucid | Digital Logic Design
Hardware-based endless runner game fully implemented on FPGA using Lucid HDL. Real-time VGA graphics, responsive controls, and obstacle generation — all in hardware, no CPU.
-
-
Exploring Nanochat
Training small LLMs on my own GPU using Andrej Karpathy's nanochat: what I learned, what broke, and what surprised me.
-
AI Combat Manikin
Building a modular AI-driven training system with Intel RealSense cameras for real-time baton tracking and responsive manikin maneuvers.
-
The boy sits down to write a poem — Costume Design
Designing and building a paper mache beast for a short film — one creature embodying Singapore's multiracial identity.
-
What The Hack 2023
Designing the visual identity for SUTD's flagship hackathon: branding, event collateral, and everything in between.
-
Pixel Run
Designing and building a hardware-based endless runner game on FPGA, fully in Lucid HDL with VGA output — no CPU involved.