Achu Mukundan https://achumukundan.dev Software engineer in Toronto building developer tools, local AI software, and audio tools. en-ca Sat, 12 Sep 2026 00:00:00 GMT Ditching llm-scaler: Qwen3.8-27B on my B70 https://achumukundan.dev/notes/intel-arc-pro-b70-inference-stack https://achumukundan.dev/notes/intel-arc-pro-b70-inference-stack Sat, 12 Sep 2026 00:00:00 GMT Local inference Leaving llm-scaler behind, patching vLLM with GPT-5.6 Sol, and getting MTP4 working. My setup as of August 20.