CUDA path tracer

Pathtracing rendering engine written from low-level primitives in CUDA. The project implements the Cook-Torrance reflectance model to achieve photorealistic physically based rendering (PBR). Standard 3D mesh formats are supported through the Assimp library, allowing models to be rendered as they would appear in Maya or Blender.

NN-guided MCTS

Pytorch pipeline to train neural network guided MCTS algorithms to play TicTacToe and Ultimate TicTacToe. Neural architectures differ from typical MCTS in that they use a neural network to estimate the outcome/pick a move using their current position. To train these networks, two copies of the model play against each other to generate training data. Each position taken is considered an input with the subsequent move taken and eventual game outcome being considered the outputs. This cycle of self-play then training can be repeated until the resulting model achieves superhuman performance.

Rubik's Cube solver

Interactive 3D Rubik's Cube and solving algorithm. This web app uses the Three.js to render a fully manipulatable Rubik's cube in the browser. The app can also communicate with a python based API server that implements the Thistlewaite solving algorithm. This allows the web app to provide step-by-step solve instructions for any Rubik's cube configuration in 52 moves or fewer.

About Me

Hello, I'm Elliot Chung, a full-stack software engineer with experience spanning robotics, embedded systems, web applications, and video games. That breadth has given me a strong ability to understand systems from multiple perspectives—from low-level hardware constraints to user-facing product experiences.

My journey into software began at 13, when I first started programming. In high school, I competed in my school's robotics team, which sparked an early interest in building systems that bridge software and hardware. I went on to study computer science at UCSD, where I gained hands-on experience in machine learning through research labwork. Today, I work as a full-stack software engineer at Franklin Wireless, an IoT and networking company, where I build integrations between web client frontends, AWS cloud infrastructure, and physical network devices.

My technical range allows me to see the bigger picture while still being comfortable diving into the details. I enjoy working at the intersection of different technologies and I'm especially motivated by work that requires connecting ideas across disciplines.

Scroll Down