https://calebgarber.com/tags/llm-serving/ Recent content in Llm-Serving on Caleb Garber Hugo -- 0.165.0 en-us Thu, 01 Oct 2026 00:00:00 +0000 https://calebgarber.com/posts/how-my-homelab-keeps-the-llama-stack-current-and-measured/ Thu, 01 Oct 2026 00:00:00 +0000 https://calebgarber.com/posts/how-my-homelab-keeps-the-llama-stack-current-and-measured/ The automation behind a local LLM stack: four daily llama.cpp source builds with atomic symlink swaps, a smoke-test gate on llama-swap releases, and the bench findings that shape how two unequal cards get served.