Graphics, game AI, and systems programmer. C++ · Rust · Vulkan · DirectX 11/12 · OpenGL · CUDA · Unreal Engine 5
I like knowing how things work all the way down, so I build them from scratch: a game engine with four graphics backends, a deep learning library with its own CUDA kernels, a game AI plugin for Unreal. Most of what's here is meant to be read as much as used.
SleakEngineC++23 engine with DirectX 11, DirectX 12, OpenGL and Vulkan behind one rendering interface. SleakCraft is a voxel sandbox built on it. |
RustVKA modular Vulkan renderer in Rust on raw ash bindings. No wrappers between me and the spec.
|
NavLabPathfinding you can watch: A*, jump point search, flow fields, and a polygon navmesh with funnel smoothing. |
| Project | What it is |
|---|---|
| Tactix | Game AI plugin for Unreal Engine 5: Utility AI, GOAP, HTN, cover, influence maps, squads and formations |
| FastNN | Deep learning library in Rust with CUDA kernels. Autograd, transformers, KV-cached generation, all from scratch. Try it in the browser |
| ViewCam | Turns your phone into a wireless webcam. Qt 6 desktop app plus an Android client |
| StealthDemo | Stealth enemy AI in UE5 with behavior trees, services and C++ tasks |
| Evo-Engine | Evolutionary computation in Rust: GA, DE, CMA-ES, NSGA-II, island models |
| raft-kv | Distributed key-value store on Raft, with gRPC and RocksDB |
| rustfm · ZipRS | A terminal file manager and a desktop archive manager, both in Rust |
- game-ai-cheatsheet: which game AI technique to use and why. FSMs, behavior trees, utility AI, GOAP, HTN, steering, A*, JPS, flow fields, navmeshes and influence maps, each with a small tested implementation.
- graphics-interview-questions: 192 graphics, rendering and engine programming questions with short answers that explain why.
FastNN is part of Hacktoberfest this year.
The issues labelled good first issue
each point at the exact file and line and say what a fix looks like, and most
don't need a GPU. Start here.
Arch + Hyprland.
