About Me
Education
- Inha University, Computer Science Engineering (Mar. 2023 – Present)
Experiences
- Google Machine Learning Bootcamp, 5th (Jul. 2024 – Nov. 2024)
- Undergraduate Research Student, Department of Nursing (Oct. 2024 – Dec. 2024)
- Undergraduate Research Student, Department of Aerospace Engineering (Jan. 2025 – Present)
OpenSource
Korean News Scraper
Apr 2024 | GitHub Repository
Developed and deployed a Python library for collecting data to train Large Language Models (LLM).
This was my first Python library, and while there were many areas to improve, it was a great experience learning about Python deployment and automation.
Notion News Crawler
Jul 2024 | GitHub Repository, Blog
Developed a crawler for collecting news by category from Notion. During development, access to Notion’s database was challenging, leading to the creation of a Python library for better integration. I set up a server running on a Raspberry Pi that works every 4 hours to upload relevant news to Notion.
Projects
Genetic Algorithm-Based Autonomous Drone Simulation
Jan 2024 – Feb 2024 | GitHub Repository
In this project, I developed a 2D drone simulator from scratch using Pygame, designed to navigate a drone to a target destination while avoiding obstacles. To discover the optimal flight path, a Genetic Algorithm was implemented to solve the control problem.
Key Implementations
- Custom Physics Engine: Instead of simple coordinate translation, I built a physics engine where the drone’s movement is realistically simulated by calculating acceleration and velocity based on the thrust from its four individual rotors.
- Pygame Environment: Developed the entire simulation environment—including the drone, obstacles, and target visuals—from scratch using the Pygame library.
- Genetic Algorithm:
- Genes: A “gene” was defined as a sequence of force commands applied to each of the drone’s rotors.
- Fitness Function: The fitness function was designed to award a higher score for closer proximity to the target, guiding the population toward the goal over successive generations.
- Evolution: More efficient flight paths were generated with each new generation by applying selection, crossover, and mutation operators.
Core Technologies
Python,Pygame,NumPy
MinGPT: An LLM Implementation Based on ‘Attention Is All You Need’
Mar 2024 – Jun 2024 | GitHub Repository
This project is a from-scratch implementation of a GPT model based on the “Attention Is All You Need” paper. The primary goal was to develop a deep, functional understanding of the Transformer architecture by translating its fundamental components into code.
Key Implementations
- Implemented the core mechanisms of the Transformer, including Self-Attention and Multi-Head Attention.
- Applied Positional Encoding to enable the model to understand the sequence and order of tokens.
- Designed the complete Transformer Block structure, combining the attention layers with feed-forward networks.
- Built the model’s core logic from scratch using fundamental PyTorch operations instead of high-level APIs like
nn.Transformerto solidify foundational knowledge.
Core Technologies
Python,PyTorch
SurvivalRL: Reinforcement Learning
Mar 2025 – Apr 2025 | GitHub Repository | Blog
This project is an ecosystem simulation where AI agents learn survival strategies. Using reinforcement learning, agents autonomously discover optimal policies to maximize their survival rate in a virtual environment by performing actions such as foraging for food and evading predators.
Core Technologies
ML-Agents,PyTorch,PPO Algorithm
Awards
- Minister of Science and ICT Award, Korea Code Fair Hackathon (2022.12)
- TOP 6 in Data Analysis, Big Data & AI Competition with AWS, KT AICE (2023.04~07)
- 3rd Place, Silver Prize, Inha University Capstone Design Competition (2023.05~10)
- Team Excellence Award, Inha University Carbon Neutral Academy 2nd Cohort (2024.06)
- Individual Excellence Award, Inha University Carbon Neutral Academy 2nd Cohort (2024.06)
- 1st Place, Grand Prize, Inha University Capstone Design Competition (2024.05~10)
- Kaggle: Top 3.7% (81st out of 2,234) in Binary Classification of Insurance Cross Selling (2024.07)