Hi

I am into |

Projects

DQN Agent for Tic-Tac-Toe

DQN Agent for Tic-Tac-Toe

Trained a Deep Q-Network (DQN) agent to play Tic-Tac-Toe using a custom Gymnasium-compatible environment. Explored self-play and reward shaping to enable learning in a deterministic, small-state game.

DQN Agent for Tic-Tac-Toe

Deep Q-Learning on FrozenLake

Deep Q-Learning on FrozenLake

Implemented Deep Q-Network (DQN) to solve the FrozenLake environment (4x4 and 8x8) using Gymnasium. Experimented with neural network sizes, map layouts, and hyperparameters to study reinforcement learning stability and convergence.

Deep Q-Learning on FrozenLake

DQN Agent for Snake

DQN Agent for Snake

Trained a Deep Q-Network (DQN) agent to play Snake using a custom Gymnasium-compatible environment. Focused on state representation, reward design, and experience replay for a partially observable grid world.

DQN Agent for Snake

Policy Gradient Agent for CartPole

Policy Gradient Agent for CartPole

Implemented a cross-entropy policy gradient agent to solve the CartPole-v1 environment using Gymnasium and PyTorch. Extended the task by training and evaluating the agent under randomized initial pole angles and positions to improve robustness and generalization.

Policy Gradient Agent for CartPole

Timeline

Deep Learning Specialization

DeepLearning.AI · Andrew Ng

Currently pursuing the Deep Learning Specialization

Feb 2026 - Present

Machine Learning Specialization

DeepLearning.AI · Andrew Ng

Completed the Course. Covered supervised and unsupervised learning, regression, classification, neural networks, and best practices for applied machine learning.

Dec 2025 - Jan 2026
2 mos

China Scholastic Competency Assessment (CSCA)

Academic

Completed the CSCA exam administered by the China Scholarship Council

Dec 2025

SAT Exam

Academic

Achieved a score of 1360 on the SAT taken on October 4, 2025

Oct 2025

IELTS Exam

Language

Achieved an overall band score of 7 on the IELTS Academic test

Aug 2025

Machine Learning Exploration

Self-Study

Began learning machine learning basics and experimenting with small projects. Gradually shifted focus away from full-stack web development toward AI, which became the primary technical interest

Feb 2025 - Present

Higher Secondary School

Academic

Completed higher secondary education and obtained the national HSC.

Dec 2024

Secondary School

Academic

Completed secondary school education and obtained the national Secondary School Certificate

Dec 2022

Stack Overflow Contributions

Q&A

Started by using Stack Overflow to solve coding issues in React, JavaScript, and HTML. Later began answering questions during the early pandemic months, eventually posting 100+ answers and earning ~4k rep

Jan 2020 - Oct 2020
10 mos

Started Programming

Self-Study

Started tinkering with HTML and basic scripting in free time as a hobby. Tried PHP for a while, deepened work with JavaScript (frontend and Node.js backend), then added Python. Focused mostly on learning full-stack web projects

Jan 2017

Skills

Python
Python
JavaScript
JavaScript
NodeJS
NodeJS
TypeScript
TypeScript
PHP
PHP
Machine Learning
Machine Learning
Deep Learning
Deep Learning
PyTorch
PyTorch
Scikit-learn
Scikit-learn
Pandas
Pandas
NumPy
NumPy
Matplotlib
Matplotlib
OpenCV
OpenCV
Data Mining
Data Mining
SQL (mysql + sqlite)
SQL (mysql + sqlite)
MongoDB
MongoDB
React
React
Next.js
Next.js
Sass
Sass
Tailwind CSS
Tailwind CSS
GraphQL
GraphQL
FastAPI
FastAPI
Git
Git
Docker
Docker
Bash
Bash
LaTeX
LaTeX

About Me

I started coding during school life — just playing with HTML and simple scripts because I thought it was magical that a computer could do something when I told it to. Back then, I had no plan, no guidance, and no idea where it would lead. I was completely aimless — just poking around, breaking things, copying code from forums, and trying to figure out why it worked (or didn't). That messy, self-driven exploration became my normal.

Over the years, I kept building things: small websites in PHP, interactive pages in JavaScript, and later full apps with React and Node.js. What fascinated me most wasn't just building things — it was taking them apart. I loved reverse engineering: reading others' code, understanding their logic, and often improving or repurposing it. That habit shaped how I learn. I've never formally studied languages like C, but if a codebase uses patterns similar to PHP or JavaScript, I can usually understand much of it — not by memorizing syntax, but by following the logic.

When school got intense, I stepped back from building full-stack web apps — but I couldn't stay away from code. I turned to data mining, which felt like the next level of reverse engineering: figuring out how websites structure data, bypassing anti-bot measures, scraping intelligently, and turning messy HTML into clean, usable datasets. It was challenging, sometimes frustrating, but deeply satisfying — like solving a puzzle that kept changing. And unlike full projects, it was doable with just a little time each day.

Then I found AI and machine learning. The first time I built a system using OpenCV, I let a user click and crop questions from a practice test image, then auto-saved and organized those crops for later review — effectively turning a static PDF into an interactive study tool. Around the same time, I watched reinforcement learning models (like those by YouTuber Yosh) learn to play games from scratch, slowly discovering strategies no human had taught them. Seeing an AI recognize patterns it had never seen before — and adapt — made everything click: this is what I've been circling my whole life. AI combines the interactivity I loved as a kid, the reverse-engineering mindset I developed over years, and the deep, open-ended challenge I've always craved.

Hence, my goal now is clear: to push the limits of what AI can do — not just by using it, but by contributing to the research that shapes its future.

Contact Me