Brandon Cui
CE & CS @ USC · SWE intern
About
I'm a software engineer and a student studying Computer Engineering & Computer Science at USC. This summer, I'll be interning as a software developer at Journal Technologies, and I'll be continuing my research on VLM reasoning at USC. When I'm not working or studying, I like to play basketball, snowboard, or play the piano and guitar.
Education
-
University of Southern California — B.S. Computer Engineering & Computer Science
2024 – 2028 · GPA 3.8/4.0
Coursework: Intro to Programming, Data Structures and Object-Oriented Design, Discrete Methods in Computer Science, Intro to Embedded Systems, Distributed Systems for the Internet of Things, Full-Stack Web Development, Calculus III, Linear Algebra and Linear Differential Equations, Probability Theory
Experience
-
Incoming Software Engineer Intern at Journal Technologies
June 2026 — Los Angeles, CA
-
Undergraduate Research Assistant at USC Viterbi
Dec 2025 – Present — VLM reasoning benchmarks
-
Lifeguard at Stanford Recreation and Wellness
Jul 2022 – Aug 2024 — Stanford, CA
Projects
-
Volt — EV Trip Planner — Full-stack EV trip planner computing optimal charging-stop routes across 2,700+ US Tesla Superchargers via a battery-aware A* graph search.
TypeScript, React, Node.js, AWS Lambda, DynamoDB, CDK
-
CS2 Skin Pricer — LightGBM quantile-regression models trained on 2,700+ live CSFloat knife listings to predict fair-value price ranges and flag underpriced skins, fed by a self-labeling Raspberry Pi pipeline.
Python, LightGBM, pandas, SQLite, Streamlit, Raspberry Pi
-
USC RateMyPlate — Full-stack dining hall rating app for USC students; Playwright scraper auto-pulls daily menus.
TypeScript, Next.js, React, Tailwind CSS, Supabase
-
IoT Posture Monitor — Two-node real-time posture system streaming webcam metrics over MQTT to a Raspberry Pi.
Python, Flask, MQTT, Raspberry Pi, OpenCV, MediaPipe
-
nanoGPT — ~10.8M param decoder-only transformer (6 blocks, 6 heads, 384 dim) trained char-level on Shakespeare.
Python, PyTorch
-
Arduino Speed Trap — Embedded vehicle speed trap on Arduino measuring pass-through speed between two sensors.
C, Arduino
Skills
- Languages — C/C++, Python, Java, JavaScript, TypeScript, HTML/CSS, SQL
- Technologies — React.js, Node.js, Express, Next.js, Flask, PyTorch, NumPy, pandas, Linux, Docker, Kubernetes
- Tools — Git, Vercel, PostgreSQL, Postman