Ji Sung Han!
And I'm a
Passionate about AI innovation and data-driven solutions. Combining statistical analysis with cutting-edge AI technologies to create impactful applications.
About Me
My journey
Currently studying Computer Science and Statistics focusing on Software Engineering and Data Science at the University of Toronto. My passion lies in bridging the gap between AI and software engineering, aiming to become a versatile developer who excels in both domains.
Driven by endless curiosity, I've explored diverse subjects beyond my core focus, believing that learning is a lifelong journey. This interdisciplinary approach has shaped my unique perspective in problem-solving and innovation.
Experience
My Professional JourneySoftware Developer, Open Source
LangChain-AI
Developed an optimized Dropbox Document Loader using TypeScript and Node.js, integrating into LangChain's AI-driven document pipeline to streamline document retrieval and processing, handling 10,000+ files efficiently.
Machine Learning Developer Intern
Chemtronics
Developed a computer vision-based ADAS perception model using YOLOv5 and DeepSORT, leveraging Darknet (C++, CUDA) for real-time vehicle and pedestrian detection, improving object detection accuracy by 20% and evaluated on 50,000+ test samples.
Software Developer Intern
Maple Bit Services
Optimized the payment processing system for the company's Point of Sale (POS) platform using Python (FastAPI), C++ (gRPC), and PostgreSQL, reducing transaction times by 35% through asynchronous processing and caching.
Projects
My Project PortfolioLifomation
Advanced AI-powered document search engine using RAG with FAISS and OpenAI Embeddings, enabling semantic search, deadline-based retrieval and contextual recommendations. Implemented OCR-based document processing with Textact, extracting key entities from scanned documents and improving search accuracy by 85%.
LearnQuest
Inclusive VR educational platform with instant translation and versatile note-taking features. Top 8 finish out of 80 teams at Hack the Valley '23. Integrated AI-powered translation using GPT API.
Resumify
Developed an AI-powered resume generator using GPT API, automating resume creation with ATS-optimized formatting and real-time feedback. Designed a scalable full-stack system with Django, MongoDB, Firebase, enabling seamless user authentication, resume storage, and efficient document generation.
Cerberus-Detection
Developed a Web Application using ipywidgets to detect AI-generated text with 93% accuracy. Implemented Deep Learning model with LSTM architectures, testing on 166,770 samples.
AI-Generated Image Detection
Ensemble Deep Learning Model integrating EfficientNet, ResNet, and Swin Transformer using Late Fusion, achieving 92% F1-score on 19,986 tested images. Implemented end-to-end training pipeline with advanced optimization techniques.
QuantML
ML-driven hybrid trading model combining technical indicators and macroeconomic factors, increasing Sharpe Ratio by 42% and achieving 14% higher Win/Loss Ratio. Implemented adaptive signal generation using Bayesian Optimization.
Autotrace
Led development of Android app for tracking license plate information. Implemented Firebase backend and Java frontend using MVP design pattern and Clean Architecture principles.
MarkScope
Web application for academic management with secure authentication and role management. Led full-stack development ensuring seamless user experience and secure data handling.
Pintos OS System
In a team of 3, developed and optimized core OS functionalities, including thread management, process management, virtual memory, and file systems, leveraging multithreading.
ConcurrentSystemStats
Engineered a high-performance concurrent system monitoring tool in C, leveraging advanced Linux system programming techniques. Implemented process forking and IPC with pipes for efficient real-time data collection.
Quotis
Architected and led the development of a real-time social platform connecting users with handymen, featuring multimedia job postings, WebSocket-based messaging, and live arrival tracking. Implemented AWS S3 for secure file storage, MongoDB for scalable data management, and a React Native, TypeScript stack for a responsive cross-platform experience, securing $4,000 in grants with 20+ freelancers waitlisted.
RainbowBreakout
Implemented the classic Breakout game in MIPS Assembly, featuring bitmap display manipulation, real-time collision detection, and keyboard input processing. Developed comprehensive game mechanics including paddle movement, ball physics, and brick destruction with multiple color layers.
Get in touch
Do you have a project in your mind, contact me hereFind Me
Email: yhs24077@gmail.com
Academic Email: jishan.han@mail.utoronto.ca
Tel: +1 (647)-321-9811