Jayden Ruddock

I started Internship Grabber because searching for internships meant jumping between job boards and keeping track of applications somewhere else. I wanted one place to find opportunities and follow an application from the first listing onward.

I built the web app and automated the collection and parsing of job listings, using AI to turn job text into structured information students can search and track.

My friend and I wanted to track what we ate at UT's dining halls. The food we were eating wasn't in the apps we used, so I started building one around our own routine, with food logging, workouts, fasting, and progress tracking.

I built it with Next.js, TypeScript, Convex, and Stack Auth. Scheduled syncs collect menus from three UT dining halls into shared storage, and a cache with expiry avoids repeating the same food API requests. I'm working toward a campus release.

I built Cody to run a local AI agent that could work with files and scripts across my devices. It connects llama.cpp and Hermes Agent with Python scripts and n8n webhooks.

I added output validation and grammar constraints to catch malformed tool calls before execution, plus timeout handling and container recovery for failed tasks.

In UT Austin's Freshman Research Initiative, I work on co-adaptive reinforcement learning for autonomous driving. We use simulated passenger comfort feedback to guide updates to trajectory policies in Python and PyTorch.

I write evaluation scripts to compare collisions, reward stability, and policy convergence across training runs, and process driving telemetry on TACC's Linux GPU clusters.

My team built SkyLore, a citizen science app that won Best Celestial Project at HackTX 2025. We put together a working app in 24 hours to help people explore and contribute sky observations.

I built Python and Streamlit views that mapped light pollution reports and observation entries on interactive heat maps.

I built a VS Code extension that uses OpenAI to explain programming errors in plain language. It gives developers a way to understand an unfamiliar error while staying in their editor, across different programming languages.

I built the official website for UT Austin's NSBE chapter from scratch. I wanted it to be useful beyond introducing the organization, so I brought attendance tracking, member points, and academic resources into one place.

The website is now part of the chapter's day-to-day operations. Officers use it to manage member records and chapter resources, with access based on their role.

I helped build Mentor Finder for students in Texas Robotics to request mentors and for program managers to review pairings. The app serves more than 60 users.

I wrote reactive Convex queries and mutations for application queues and mentor capacity, and set up Clerk authentication and admin dashboards for matching and rosters.