Implementing with AI
I have a strong passion for learning and implementing new things. My interest in AI, which began with robot soccer, naturally extended to IoT sensor processing and machine learning models. Recently, I have been focusing on LLMs using LangChain, working on personal projects in this area. As new technologies continue to emerge, I am particularly interested in leveraging these advancements to create innovative business solutions.
Private Chatbot
Role: Personal Project
As a developer, I had the opportunity to create a private chatbot application from scratch. This project involved developing a chatbot that allows users to upload personal documents for Q&A interactions, so I no longer need to read and study documents manually. To enhance my personal skills, I also implemented functionality to generate quizzes from JSON-formatted data. Additionally, I integrated OpenAI's GPT-API for language model interactions and implemented a deployable LLM using Ollama for local machine learning model deployment.
Technologies Used
Development : LangChain, Streamlit
API : GPT-API, HuggingFace
Local Deployment : Ollama
Soccer Robot
Role: Software Engineer
As a founding member, I had the unique opportunity to be involved from the very start with the Nao robot project. Training AI for robots requires substantial data. Since teaching real robot movements is too time-consuming, I leveraged the Pygame library to create simulation games, using the data generated to train the robots effectively.
Technologies Used
Perception : ball/field recognition (OpenCV)
Planning : optimize path (Sim2Real, stablebaseline)
Control : robot movements (C++, B-human)
Secured 3rd place at RoboCup 2023 in France
IoT Control system
: From sensor to the Web
Role: Hobby
As a hobby, I engage in Raspberry projects. I recently purchased an BME280 chip, soldered it, and developed a simple web system that processes data in the cloud. This setup allows me to manage my schedule and monitor the temperature and humidity in my room.
Technologies Used
H/W : Raspberry-Pi, BME280(I²C)
System : Python Flask