About

hi, i'm richard

I'm a 3rd year computer science student at Purdue University. I'm based around Indiana and New York City, but I'll go wherever life takes me. I love expressing my creativity by incorporating my vision in software. Whether that be fullstack or AI systems, I'm always willing to learn more.

Some of my hobbies include photography, bouldering, editing, music, trying new restaurants, and reading manwha.

Connect with me

Projects

creative outlets implemented

Grass Hopper

2D game based on movement of character through shifting foot pivots in a circular motion. WIP

RenderFlow github demo video

A performance profiler for DaVinci Resolve. Point it at an open project and it works out which clips are slow to decode and which are expensive to render, then drops markers on the timeline right where the trouble is. It can also fix what it finds: generate lightweight proxies for the heavy clips, turn on smart caching where effects are piling up, and undo all of it if you change your mind. Works on the free edition too, through a small bridge script. Written in Python on top of FFmpeg and Resolve's scripting API.

Sift demo video github

A YouTube tool that recommends channels instead of videos. You tell it what you like to watch and it digs up smaller creators making that kind of thing, the ones the normal algorithm never shows you. You can also hand it a channel you already like and it finds similar ones. Sign in with Google and it pulls in your real subscriptions too. Made with Node and React.

Sift landing page: Find the creators the algorithm buries.

MeetHere site github

An upgraded take on When2Meet, built with a couple teammates for seamless meeting creation. Everyone marks their availability and it finds the best overlap, and it also calls the Google Maps API to suggest a fair meeting spot based on everyone's location. I designed a full UI/UX refresh in Figma and Google Stitch, but unfortunately haven't had the chance to actually implement it yet, so the live site is still running the old interface for now. Built with Node.js, Express.js, MongoDB, and JWT, deployed on Vercel and Render.

Question Classification slides github

Worked on a machine learning tool with three teammates, mentored by a Google engineer, that recommends questions to students across three science disciplines: biology, chemistry, and physics. I built the website that hosts it and scraped data from textbooks and Reddit threads. We presented it to a few hundred students and faculty at the end of the program.

This Site

Just one HTML file, no build step. The landing page is a small 3D scene built with Three.js. You can drag the camera around, fade between day and night, or click it to walk through the scene and take photos that end up in the Gallery. Everything else sits on top like a paper booklet. It runs on GitHub Pages.