><((('>
hi, i'm @matttwh0
Matthew Tran
cs @ sdsu · full-stack + ai
About
Hey, thanks for visiting my website! My name is Matthew and I'm a developer from SDSU. I build full-stack apps and AI agents, and I've shipped production code in fast-moving startups. At NXP, my team and I built an agent that generates diagrams to speed up chip design and debugging.
In my free time, I like to help my local community by building websites for their businesses. I also love fly fishing and playing basketball at my local recreational center.
Stack
PythonTypeScriptSQLReactNext.jsFastAPINode.jsLangGraphAWSPostgreSQLROS2Linux
Experience
Software Engineering Intern
feb – may 2026
Susty
- Built and deployed the company's web application on Vercel, owning the full development lifecycle from design to launch.
- Launched a production web app from scratch by translating Figma designs into Next.js.
- Improved on-page SEO score by 33% with XML sitemaps and robots.txt.
- Collaborated directly with the CEO to align product direction and ship features on a startup timeline.
Cloud AI Engineering Fellow
feb – apr 2026
NXP × CTRL SDSU
- Architected a multi-agent AI workflow with LangGraph on a team of 5 to generate chip block diagrams from RTL code.
- Developed a unit & integration test suite for AI agents to catch failed JSON schemas and drive LLM self-correction.
- Wired the AWS Bedrock API, IAM roles, token usage, and cost tracking across models.
- Built a Graphviz diagram pipeline with hierarchical node abstraction, plus a local interactive visualization server for engineers.
Software Engineering Intern
oct 2025 – jan 2026
Dynamic Active
- Designed and implemented the FastAPI backend of a Learning Management System from the ground up.
- Shipped 14+ RESTful APIs powering auth, assignments, submissions, and event workflows.
- Designed the PostgreSQL schema for users, assignments, submissions, grading, and event tracking.
Robotics Perception Intern
feb 2025 – now
Real-Time Embedded & Control Systems Lab · SDSU
- Improved vehicle object detection to 92% mAP50-95 at 85 ms average latency, reducing crash rate.
- Integrated a YOLOv8-nano model trained on 833 custom-labeled images into an autonomous F1Tenth racing platform's Intel RealSense setup for real-time localization and obstacle detection.
- Developed a ROS2 node that dynamically crops and resizes the camera feed to cut noise and boost performance.
- Integrated the CARLA simulator with the F1Tenth platform for faster perception testing and validation.
Machine Learning Research Assistant
sep 2024 – aug 2025
AI4Business Lab · SDSU
- Developed automatic pavement distress detection algorithms with CNNs on a research team.
- Processed 1,000+ images using a weakly supervised learning approach.
- Trained distillation models on high-performance remote servers with large pavement distress datasets.
Information Technology Intern
jun – aug 2023
Lavner Education · University of San Diego
- Instructed cohorts of 10–13 students in Lua, robotics, and object-oriented programming.
- Developed streamlined curriculums spanning computer science, robotics, and game development.
AACE ↗
Agentic Abstraction for Chip Engineering, a multi-agent pipeline that turns RTL code into clean, single-page chip block diagrams.
3Dify ↗
Turns short videos into 3D-print-ready meshes via gaussian splatting. Led a team of 3, React UI, Firebase auth, Three.js preview.
Susty ↗
Marketing site for a consumer-first sustainability ratings platform, built and shipped from Figma designs to production on Vercel.
KL Milky Nails ↗
Website for a nail salon in Encino, CA, designed and built end-to-end with Claude, from art direction to deployed site.
Education
San Diego State University, B.S. Computer Science
aug 2022 – may 2026