Chart the northern passes and log what you find there.
Click anywhere to start

Hello, Welcome to my website

I venture into areas like
competitive programming
machine learning
game development
web development
competitive programming
machine learning
game development
web development
Honkai Star Rail

About Zhaoyi

I mainly take part in competitive programming. I have competed in programming competitions, such as the National Olympiad in Informatics. I'm rated 1642 on Codeforces.

I'm currently working on an AI Research Agent Chatbot that provides systematic steps on how to tackle its prompt questions.

I'm very enthusiastic in the field of competitive programming and artificial intelligence. I plan to win more awards in these 2 areas and develop my current project.

I've been playing Honkai Star Rail since version 1.4

Skills

Image

Competitive Programming

I practice competitive programming almost daily and participated in various programming competitions, such as codeforces division rounds.

Image

Training

I have been a lecturer in CPIC 2025 and NUS Dec Course 2025

I really enjoyed setting contests and planning lectures during that time

Image

Web Development

I have just started to gain interest in web development and created my first website.

I would like to learn more about front-end web development in the future.

Image

Mathematics

Mathematics and Computing is inseparable.

Math highlights certain patterns and ways to compute a certain value more efficiently.

Computing makes the computations come true.

Image

Machine Learning

In National Youth Tech Championship, My team coded an autonomous drone and let it navigate through obstacles through machine learning.

Image

Game Development

I have made some games with Swift Playground.

They can be downloaded in the portfolio below

Image

Gamer

I have managed to build my hsr characters to considerable high levels

Projects

Multi-agent Research System

I have been wondering how the multi-agent systems were built for research tasks, such as the deep research function provided by the companies like Anthropic or OpenAI. After studying a lot of references, I decided to kickstart this project in order to understand the internal architecture and workflow of multi-agent research systems.

In my current Version 1 code: When a user submits a query, (1) the manager agent analyzes it, plans research steps, and then delegates to subagents to explore the various steps. (2) The subagents search across the web by iteratively generating Python code to call the search tools, and (3) then returning a list of gathered information to the manager agent so it can compile a final answer.

Swift Games
Please download Swift Playgrounds to view the Swift Games
post-img
Quiz
post-img
Wordle
post-img
Todoist

Portfolio

My Honkai Star Rail builds

Contacts

Please click on mini logo icons in the blue box to be redirected to my accounts