Joshua Lam

I've built and deployed full-stack web applications at U Points Limited using Next.js, TypeScript, Tailwind CSS, PostgreSQL, Prisma ORM, and RESTful APIs.

At PolyU's Department of Applied Social Sciences, I delivered an educational web game end-to-end, extending RPG Maker MZ with custom JavaScript and iterating through educator-led testing before HTML5 deployment on itch.io. I also built an AWS-backed workflow for capturing student choices, synchronizing progress, and analyzing learning outcomes with Python.

My projects apply AI to practical products, including a WebAR campus landmark guide using YOLOv8n, FastAPI, an LLM chatbot, and Docker hosting. I've also developed queue-counting and wait-time estimation with YOLOv8 and OpenCV, reproduced a binocular vision model in TensorFlow, and optimized house-price prediction models with XGBoost, LightGBM, and ensemble learning.

I'm especially interested in turning machine learning, computer vision, and LLM capabilities into useful, deployed experiences.