Projects
Here are some of my latest projects and works I have completed recently. First, I wanted to introduce the AI project I did with my classmates for my capstone project.
Gendia – Multimodal AI Generation Platform
Project Manager & Full-Stack Developer
2025.06 – Present
Gendia is a multimodal AI platform that allows users to generate images, videos, and audio through a unified and intuitive interface. As the project manager, I lead the planning, feature design, and overall development direction, ensuring the product evolves with clear structure and practical user workflows.
On the technical side, I designed the backend architecture and built the core API system that connects the platform with various AI services. I also handled Docker-based deployment and GPU inference setup, ensuring stable and efficient model execution. While collaborating with frontend developers, I focused on integrating backend features and maintaining smooth communication across all parts of the system.
Gendia is an ongoing project aimed at delivering a streamlined, accessible creation experience where users can easily work with multiple AI-generated media types in one place.
AniCreator – AI Animation Generator Website
Full-Stack Developer
Oct 2025 – Nov 2025
AniCreator is an AI-powered demo website that generates stories, scenes, and short video clips from a user’s idea or theme. The platform connects story generation with automated scene creation and video assembly, offering a seamless demonstration of an end-to-end AI storytelling pipeline.
For this project, I designed the overall structure and built the core components across the backend, AI integration, and user flow. I implemented the main pipeline logic, connected the required AI services, and developed the foundational design, frontend, and backend elements needed to present the demo as a complete, functional prototype.
Jinza – Skincare Chatbot & Clinic Price Finder
AI & Backend Developer
Jul 2025 – Nov 2025
Jinza is an AI-powered skincare recommendation service that suggests suitable products and compares dermatology clinic prices. I built the backend and AI logic behind the personalization flows, including a RAG-based system that analyzes user skin concerns and produces tailored product suggestions through structured retrieval and agent-driven reasoning. I also developed the authentication system, API structure, and the data pipeline that crawls and organizes clinic pricing information, enabling accurate recommendations and transparent cost comparisons for users.
MemeCleaner – AI Meme Text Cleaning & Translation Tool
Team Leader, Full-Stack Developer
Apr 2025 – Jun 2025
MemeCleaner is an AI-based web tool that removes or translates text in meme images through a fully automated pipeline. The system detects text regions using OCR, extracts and reviews the bounding boxes, and performs inpainting to cleanly remove the original text. It then generates either a translated or filtered version of the content using an LLM and places the new text back onto the meme with the correct font and layout. I developed the entire platform end-to-end, including the full pipeline logic, backend processing, inpainting integration, and the complete frontend interface.
.png)
EduPrompt – AI-Based Prompt Learning Platform
Team Lead & AI/Backend Developer
Jan 2025 – Jun 2025
EduPrompt is an AI-powered learning platform where teachers create image-based quizzes and students try to generate matching images by writing prompts. The system evaluates similarity and provides automated feedback, helping students learn prompt writing in a fun and interactive way. As the team leader, I planned the project structure, designed the wireframes, and coordinated the overall development direction. I also built the AI deployment pipeline, implemented backend logic for the quiz and evaluation system, and integrated the LLM/VLM-based scoring and feedback components that form the core of the platform.
AI-Based Voice Recognition Coffee Kiosk
Team Leader, AI model developping
The AI-Based Voice Recognition Coffee Kiosk is an innovative solution designed to simplify and enhance the coffee ordering process. This project leverages state-of-the-art AI models, including the LLaMA 3 language model, to provide accurate and natural voice interactions in Korean. By fine-tuning the model with over 135,000 conversational datasets, the system ensures smooth and user-friendly order management, even handling complex requests and regional dialects.
Key features include real-time voice recognition, dynamic menu updates, and advanced functionalities like modifying orders and adding items through natural language commands. Developed using Python, OpenAI GPT, and Hugging Face frameworks, this kiosk demonstrates seamless integration of speech-to-text (STT) technology and AI-driven conversational interfaces. The project emphasizes scalability, efficient processing, and adaptability to various coffee shop environments.
This system was created with a focus on improving accessibility, convenience, and overall user experience, setting a new standard for smart retail solutions.
Team Leader, AI model developping
Article: LLaMA vs. RASA in Voice Recognition Kiosks
As part of the project, I co-authored an article comparing LLaMA and RASA for voice recognition kiosks. Our paper evaluates their effectiveness in processing complex commands, dynamic language inputs, and nuanced modifications. The findings highlight finetuned LLaMA’s superior adaptability and accuracy, showcasing its potential to revolutionize customer-facing AI systems in real-world environments.
Our work was recognized with the Excellence Award (논문 우수상) at the UCWIT 2024 conference, marking a significant achievement for our research.
GET IT Ideathon & Hackathon
I have been a member of GET IT for two years, a startup club focused on innovation and entrepreneurship. Each semester, I collaborated with teams to develop creative and impactful projects. Below are some key projects I contributed:
Refrigerator Food Management Service
UI/UX Design
This project aimed to reduce food waste by tracking and managing stored food items. The application features alerts for storage start dates and expiration periods, along with food-sharing and recipe suggestions. My primary contribution was designing the intuitive UI/UX, ensuring a seamless user experience and visually appealing interface to encourage practical usage.
PUT IT ON: Virtual Fitting Application
UI/UX Design
This project focused on developing a 3D-based virtual fitting platform to help users visualize how clothes would look and fit when shopping online. The app suggest advanced 3D modeling tools and interactive features, enhancing the online shopping experience by providing realistic clothing simulations. My primary contribution was designing the intuitive and user-friendly UI/UX, ensuring seamless navigation and an engaging user interface.
Delivery Market Service
Front-end, Web design
This project aimed to connect people needing assistance with delivery tasks to those willing to offer their help, creating a mutually beneficial platform. My role focused on developing the front-end and designing the user interface, ensuring an intuitive, responsive, and visually appealing experience for users. The design emphasized seamless navigation and usability to cater to a wide range of users, effectively bringing the concept to life.
OTHER ACTIVITIES
Bridging the Gap: Women's Workforce Participation Platform
Idea Creation, UN Forum
This project, presented at the 7th Youth Forum on SDG Implementation, proposed a technology-driven solution to address the gender gap in workforce participation in South Asia. We designed a conceptual platform aimed at empowering women through personalized job matching, self-development courses, and entrepreneurial resources.
As a finalist team, we shared our vision of how technology can create scalable and impactful solutions for societal challenges, emphasizing inclusivity and accessibility. The project highlighted the potential of apps and platforms to foster gender equality and economic empowerment.
