LLMs Explained Like System Design.

Milvus splits a vector database into stateless proxies, coordinators, specialised worker nodes and durable storage. Here is how inserts and searches flow through it, and why that design decides how it scales and fails.

Milvus segment compaction, etcd revision compaction and etcd defrag are three different operations that get mixed up during a NOSPACE incident. Here is what each reclaims, and what to do when none of them helps.

In Milvus 2.5, the write stream feeds the query layer too. A bulk insert can push up search latency with flat search traffic, and the obvious fixes (more DataNodes, more replicas, weaker consistency) often miss. Here is the mechanism and what actually works.

AI is not just automating jobs, it is creating entirely new ones, and reshaping almost every job that already exists. Here is a practical map of where the opportunities are, whether you write code or not.

Skills are not prompts. They are reusable, invokable workflows that encode expertise once and execute reliably. Learn how to design Claude Skills with the right fields, forking strategy, tool permissions, and composition patterns.

MCP is the open protocol that standardizes how LLMs connect to external tools and data. This deep-dive covers the architecture, primitives, transport layers, and what MCP is not, for engineers building with AI.