Computer Science student focused on backend engineering, full-stack systems, and AI-powered applications.

I enjoy building software that scales, makes sense architecturally, and solves real problems. I'm especially interested in APIs, system design, and applied AI — and I like shipping projects that people can actually use. Most recently, I took a project all the way from code to a live, multi-user product deployed on AWS with HTTPS.

Outside of coding, I spend time at the gym, studying hard, and sharpening my problem-solving skills.

• 🏫 Pursuing a B.S. in Computer Science at the University of California, Irvine (Expected June 2028)
• 🚀 Recently shipped a live app to AWS — Docker + Nginx + gunicorn + Let's Encrypt HTTPS on a custom domain
• 🔭 Building full-stack web apps with Flask, MongoDB, and Docker — and deploying them to the cloud
• 🤖 Exploring applied AI / LLMs — structured output, prompt engineering, and closing the loop between data and generation
• 🌱 Always learning: system design, cloud & DevOps, C/C++ internals, and clean, extensible architecture

Languages

Web & Data

AI / LLM

Cloud & DevOps

Tools

Flask · MongoDB · Docker · AWS · Google Gemini

A live, multi-user web app where users practice role-based interview questions and receive structured AI feedback, with score trends tracked session-over-session.

• Deployed to AWS EC2 as a Dockerized stack — an Nginx reverse proxy, gunicorn, and a Let's Encrypt HTTPS certificate on a custom domain
• Accounts with bring-your-own-key: email/password auth (Werkzeug-hashed passwords) and per-user Gemini keys encrypted at rest (Fernet); every endpoint scoped so users see only their own data
• Integrated Google Gemini with Pydantic response_schema enforcement to guarantee rubric-scored JSON output (clarity, structure, depth, specificity) instead of free-form text
• Engineered a MongoDB aggregation ($lookup + $group) that surfaces a user's weakest categories and feeds them back into question generation — a closed performance-improvement loop
• Wrote 26 pytest tests against an isolated DB, stubbing Gemini with an autouse fixture for fast, deterministic, offline runs

🔗 Live: jinwooworks.com · Code: AI-Mock-Interview-Coach

Flask · MongoDB · Docker

A full-stack app that auto-computes ACP brevet control times as the user types.

• Python Flask backend with an AJAX frontend for live, keystroke-driven calculation
• MongoDB persistence with submit/display functionality, deployed as a multi-container Docker Compose stack

C++

• Implemented a binary search tree from scratch using classes and composition (insert, search, delete, transplant)
• Built a student-records database on top using inheritance — verified zero memory leaks with Valgrind

C++

• Designed five ciphers (substitution, Caesar, ROT13, running key, Vigenère) using a class inheritance hierarchy
• Applied OOP design — inheritance and polymorphism — for clean, extensible encryption/decryption