Jun Xia

I like to build scalable systems with reliable infra 🧠🤖💥

Summer 2024
I was the lead developer for QC-Devs during Google Summer of Code, where I spearheaded the transformation of a specialized Python library called Selector into an accessible tool for the scientific community. This included designing an intuitive web interface using Streamlit, implementing robust data processing capabilities, and establishing automated deployment pipelines. Today, the platform enables thousands of medical chemists with minimal programming background to perform complex chemical space analysis.
GSoC Completion Certificate
2021 - 2025
BSc at the University of California, Irvine where I worked with Chen Li on query processing and optimization. This is where I first got into building open-source systems 😊

I was the tech lead at SQLRewriter, where I directed a team of 4 developers. This platform enhances the QueryBooster framework by facilitating community-driven SQL query optimization and discussion. This includes full-stack development, database management, the science of making it work, and deployment in production running on our custom lab server. Today, SQLRewriter increases the efficiency and precision of SQL query optimization, but the team's goal is to develop and deploy self-optimizing query systems that automatically adapt to changing database workloads.
SQLRewriter
Texera
I was the Teaching Assistant for ICS 31: Introduction to Programming, ICS 33: Intermediate Programming, and CS 122B: Project in Databases and Web Applications, where I was responsible for leading lab sessions, grading assignments, and designing course materials for over 100 students.

I was also a student developer for ICS Student Council where I participated in the design and development of AntAlmanac, the most popular course planning platform used by more than 35000 students at UCI.
pet projects
ICS Search Engine is a specialized search engine that indexes and enables searching across UC Irvine websites. The engine uses TF-IDF scoring combined with PageRank and implements proximity-based scoring for multi-word queries.
Fablix is full-stack e-commerce website that enables the users to search, browse and purchase movies.. The front end uses HTML, CSS, and JavaScript. The back end uses Java Servlet and MySQL for database management..
PulsePost is a full-stack platform that allows the user to login and share posts. The frontend uses Angular.js. The backend uses Express.js, Node.js, and MongoDB for database management.
Traffic-Vision-Learner is a neural network to classify road signs based on German Traffic Recognition Benchmark dataset.
BREAKOUT-GAME is a game that implements a real-time scoreboard for users who control a paddle to destroy bricks.


Stolen from Andrej Karpathy