Skills
Languages
Front end
Back end & data
Engineering practice
AI & automation
Jordan Chen
About Me
I'm a student at the University of California San Diego, La Jolla, CA (expected June 2028) pursuing a B.S. in Mathematics Computer Science with minors in Data Science and Cognitive Science (GPA 3.9). Relevant coursework includes Advanced Data Structures (C++), Systems Programming (C, GDB), and Numerical Linear Algebra.
I work at the intersection of full stack development, AI tooling, and automation, building LLM powered web apps and backend services, with interests in human computer interaction, computer vision, and statistical modeling.
Experience
Allied Legacy Partners
CurrentBuilt web dashboards and UI that use LLMs to turn meeting transcripts into action items for the team. Designed, developed, and deployed a Python backend service (Slack bot) that automates calendar, email, and file tasks. Turned free form user requests into JSON commands via LLM APIs, each routed to a Playwright automation, and raised reliability by validating every LLM generated command against a schema before it could touch live systems. Migrated legacy data to PostgreSQL (Supabase), redesigning schemas to speed up queries across 5+ systems.
UC San Diego, Halıcıoğlu Data Science Institute
CurrentHelp students debug Python one on one, grade assignments, and help lead discussion sections. Lead a study and social group for DSC 10 students, reviewing course material and building community.
Unicircle
CurrentBuilt full stack recruitment management features in Next.js, supporting applicant tracking, interview workflows, and organizational recruitment operations. Engineered candidate filtering and search functionality to streamline applicant review and improve recruiter decision making efficiency. Collaborated with a cross functional development team using Git version control to deliver scalable web application features in a production environment.
UCSD Foundation Interface Lab
Built the AI chatbot system for Meridian, the lab's npm package that adds AI analytics to any web app. Engineered a prompt to JSON schema layer so the AI renders the best fit front end UI per request. Tested UI generation and data binding end to end in a React, Next.js, and TypeScript app before each release. Worked with product designers and engineers in design meetings to turn user needs into shipped Meridian features.
Sanford Burnham Prebys
Built Next.js web apps that made data input and output intuitive for researchers, translating lab requirements into UI. Halved data prep time with a web tool converting raw lab output into ready to analyze GraphPad Prism datasets. Cut neural signal analysis time over 6x (1 hour to under 10 minutes) with automated Python pipelines on gigabytes of data.
NGYL
Created the NGYL website with HTML, CSS, and JavaScript to increase organizational awareness, roughly double volunteer sign ups, and provide a communication platform for organization missions and events. Organized and invited four industry professionals for a STEM career seminar with over 200 participants. Taught 300+ students in Python, drone robotics, Lego robotics, and machine learning.
Projects
No Code EDA
- Built a full stack AI analytics web app to explore datasets up to 25 GB and create charts in plain English.
- Designed REST APIs in FastAPI that let the React front end run data analysis workflows on demand.
- Scaled to concurrent users by containerizing each workspace with Docker in an isolated, reproducible environment.
- Redesigned the data analysis workflow with AI assisted human computer interaction (HCI) principles, letting users sort, filter, reorder, and visualize datasets through natural language.
- Integrated the Meridian package (UCSD developer package) to data bind CSV data with visualizations.
Ecommerce Dashboard
- Shipped a React/TypeScript dashboard with NextAuth login, deployed on Vercel.
- Synced per user data across devices via Firestore (NoSQL), using debounced writes to cut redundant saves.
- Brought spreadsheet style workflows to a mobile first experience, with product, order, and reporting views tailored for small business style operations.
Add image at /public/project-placeholder.jpg
DeckScan & Wash Shuffle Research
- Tripled data collection efficiency with a web app that reads deck orderings from photos (with Prof. Peter Chi).
- Connected Next.js image capture to a Python batch inference backend with CSV export across 600+ trials.
- Generated large scale datasets from wash shuffle trials for probabilistic modeling and statistical analysis.