MIMihir Inamdarinaihive.hashnode.dev·Sep 27 · 5 min readAI coding agents write bash. Windows shells don't speak it.The problem Claude Code, Cursor, Windsurf, all of them learned how to use a shell from the internet, and the internet runs on Linux. So when your agent needs to list files, it writes ls -la. It does n00
MIMihir Inamdarinaihive.hashnode.dev·Sep 24 · 9 min readI embedded all of Qdrant's docs with FastEmbed. The API would have cost 2 cents.If you're building search on top of Qdrant, you hit this question early: do you embed your documents on your own machine, or pay OpenAI to do it? Running it locally sounds like the thrifty choice. No 00
MIMihir Inamdarinaihive.hashnode.dev·Sep 23 · 9 min readOne payload index per key: what 3,685 GitHub topics did to my Qdrant diskQdrant has a tutorial called Indexing Payloads of Random Shape. It's about collections where the payload keys are themselves data. Think of a product catalog where every vendor sends its own fields, o00
MIMihir Inamdarinaihive.hashnode.dev·Jul 12 · 25 min readNatural Language as a Query Interface: qdrant-interpreter in Production Agentic RAGHow a natural language interpretation layer on top of Qdrant's vector search engine fills a critical gap in production agentic RAG systems — covering filter translation, schema introspection, architec00
MIMihir Inamdarinaihive.hashnode.dev·Jul 12 · 12 min readThe Retrieval Granularity Problem in RAGWhy indexing at a fixed chunk size breaks RAG, and how a multi-resolution index in Qdrant fixes it by deferring the granularity decision to query time. Every RAG system makes a silent assumption at i00
