Khoi Nguyen

I'm an undergraduate student at Eindhoven University of Technology studying Computer Science and Engineering, where I'm also doing Honors Artificial Intelligence and being Teaching Assistant for several courses.

Outside of university, I'm working as Junior Researcher at Paper Analyzer Research Lab, Jetbrains, where I do research on post-training and test-time scaling.

In the weekend, you might find me playing table tennis, working out, or thinking about life (the answer might not be 42 🥀).

Email  /  Scholar  /  Github  /  Linkedin  /  Blogs

Research

My research interest at the moment is mostly Reinforcement Learning. However, I'm also interested in the application of Deep Learning in embryo/healthcare and safety/alignment. My goal is for my research to solve (niche/everyday) problems and improve daily life.

A cooperative multi-agent reinforcement learning benchmark in complex, egocentric environments.

Projects

A 'Moneyball' approach to student housing. This browser extension analyzes housing probability using Z-scores and IQR outlier detection, helping users predict their chances of securing a spot for a place on Vestide.

Chaos, co-op, and high voltage. A top-down survival game where two players are physically linked by an electric tether. Coordination is key - use the tether to shield, zap enemies, or accidentally drag your partner into doom.

Modernizing nature conservation. A mobile solution for forest rangers to track biodiversity, utilizing YOLOv11 for real-time species detection and GPS-based mapping to monitor population trends.

A smart assistant for Sujiko number puzzles. Features an interactive GUI to create, edit, and instantly solve complex grids using backtracking algorithms.

A 'Trust Me Bro' C++ code formatter. Born from frustration, it forces code into compliance. Simple, aggressive, and effective.

Canvas LMS submission and comment monitor - Chirp when there's new activity (e.g. announce user when someone makes a comment on their submission; because of how Canvas is designed, this only ever appears as compact phone notification instead of email, and must use SpeedGrader to use. Too much hassle. With CanvasChirp this is done automatically if Python script is running).

Many professors hide the exam few days after results are out, and you can't view it anymore. Maybe you can download the PDF but for ANS-based exams that is not possible and requires a bunch of screenshot. This tool downloads everything (source code) of your exam, and you can view it locally, while it will look the same as when you view it on the website. This bypasses all download prevention from ANS.

Adopted from jonbarron's website.