voidwest

systems, ml, and the space where they overlap.

github ember writeup

systems & embedded

ember

cpu-first llm inference engine in rust. gguf loader, gpt-2 transformer with kv cache, zero framework dependencies. ~2,200 lines.

rustinferenceggufcpu

mini-os

minimal x86_64 kernel in rust. gdt, idt, paging, heap with 3 allocation strategies, cooperative async executor, ramdisk, and a shell with 10 commands. runs in qemu.

rustkernelx86_64no_std

ml & inference

devdocs-v2

local rag pipeline for querying pdf documents. sentence-transformers for embeddings, chromadb for vector storage, ollama for inference. fastapi backend with a vanilla html/css/js chat ui. fully local, no external apis.

pythonragchromadbollamafastapi