~/projects
Deep-dives with status, metrics, and update logs.
Crawler + Sentiment Pipeline
pausedJun 30, 2026
A polite, crash-resumable web crawler in Go feeding a hybrid sentiment-analysis pipeline in Python, deterministic signals refined by a pluggable LLM, every output traceable to verbatim source text and gated by a human-review loop.
GoPythonWeb CrawlerNLPPipelines
Mini Inference Engine
in progressJun 30, 2026
A from-scratch LLM inference serving core in C++20/CUDA including a paged KV-cache allocator, a continuous-batching scheduler, and paged attention. Built to implement the load-bearing parts of a modern engine rather than import them.
C++CUDASystemsLLMInference
This Site (kobeyoung.net)
shippedJun 30, 2026
A portfolio, technical blog, and live self-hosted LLM demo built as a monorepo with a Next.js front end, a zero-dependency Go backend, and layered abuse defenses, self-hosted on a single Hetzner box.
Next.jsGoMDXSelf-hostedSecurity