Neural Systems Dispatchhttps://neural-systems-journal.lilaceagle1.chatgpt.siteEducational sample articles from a personal engineering blog.en-usZero-Overhead Speculative Decoding with CUDA Graphs and Tritonhttps://neural-systems-journal.lilaceagle1.chatgpt.site/#lead-storyneural-systems-note-001A teaching example exploring draft tokens, verification, and GPU execution graphs. All benchmark values are illustrative.Building a Sub-50ms RAG Engine with Rust, Qdrant, and vLLMhttps://neural-systems-journal.lilaceagle1.chatgpt.site/#articlesneural-systems-note-004A practical guide to thinking about retrieval pipeline latency budgets.