Deep architectural dissections of mixture-of-experts, local LLM serving, agentic protocols, and low-latency inference kernels.https://ai.ondevkit.in/en-ushttps://ai.ondevkit.in/blog/deepseek-v4-1-flash-serving-guide/https://ai.ondevkit.in/blog/deepseek-v4-1-flash-serving-guide/Architecture breakdown of DeepSeek's October 1017 update: 1M context window, native FP8 precision, DeepSWE 77.0% benchmarks, and production vLLM and SGLang deployment flags.Thu, 22 Oct 2026 00:00:00 GMTvLLM-ServingDeepSeekvLLMSGLangBenchmarksHardwareInferenceMohanhttps://ai.ondevkit.in/blog/production-mcp-agents-typescript-guide/https://ai.ondevkit.in/blog/production-mcp-agents-typescript-guide/A comprehensive engineering guide to building Model Context Protocol (MCP) servers and autonomous agents with TypeScript, Zod schema validation, SQLite persistent memory, and host integrations.Wed, 21 Oct 2026 00:00:00 GMTAgentsMCPModelContextProtocolTypeScriptAI AgentsClaudeCursorMohanhttps://ai.ondevkit.in/blog/self-hosting-deepseek-r1-vllm-guide/https://ai.ondevkit.in/blog/self-hosting-deepseek-r1-vllm-guide/Exact memory calculations, tokens/sec benchmarks, and vLLM configuration flags for serving DeepSeek-R1 distilled models on RTX 4090, RTX 3090, and Apple Silicon.Wed, 21 Oct 2026 00:00:00 GMTLLMsDeepSeek-R1vLLMInferenceRTX4090AWQSelfHostingMohan