---
title: "Mihir Inamdar (@inamdarmihir) | Hashnode"
description: "Mihir Inamdar (@inamdarmihir) writes on Hashnode. Read their latest articles and follow their work."
url: "https://hashnode.com/@inamdarmihir"
---

@inamdarmihir

# Mihir Inamdar

@inamdarmihir · Pune, India · Joined July 2026

### [AI coding agents write bash. Windows shells don't speak it.](https://aihive.hashnode.dev/ai-coding-agents-write-bash-windows-shells-don-t-speak-it?utm_source=hashnode&utm_medium=feed)

ShellSage intercepts your AI coding agent's bash commands and rewrites them as PowerShell before the shell sees them. Local, fast, no API key.

5 min read · [aihive.hashnode.dev](https://aihive.hashnode.dev/ai-coding-agents-write-bash-windows-shells-don-t-speak-it?utm_source=hashnode&utm_medium=feed)

[Sep 27](https://hashnode.com/posts/ai-coding-agents-write-bash-windows-shells-don-t-speak-it/6ab90ddf961952e7e47a4849)

### [I embedded all of Qdrant's docs with FastEmbed. The API would have cost 2 cents.](https://aihive.hashnode.dev/fastembed-vs-openai-embedding-cost-qdrant-docs?utm_source=hashnode&utm_medium=feed)

5,084 chunks, 1.2 million tokens, 24 minutes on a 2 vCPU box. What running embeddings locally next to Qdrant actually costs, with every number labeled.

9 min read · [aihive.hashnode.dev](https://aihive.hashnode.dev/fastembed-vs-openai-embedding-cost-qdrant-docs?utm_source=hashnode&utm_medium=feed)

[Sep 24](https://hashnode.com/posts/fastembed-vs-openai-embedding-cost-qdrant-docs/6ab49039726418bbd9da7867)

### [One payload index per key: what 3,685 GitHub topics did to my Qdrant disk](https://aihive.hashnode.dev/qdrant-payload-index-per-key-98x-disk?utm_source=hashnode&utm_medium=feed)

I reproduced a Qdrant tutorial's warning on a real server, hit two walls on the way, and wrote a small CLI that spots the problem early.

9 min read · [aihive.hashnode.dev](https://aihive.hashnode.dev/qdrant-payload-index-per-key-98x-disk?utm_source=hashnode&utm_medium=feed)

[Sep 23](https://hashnode.com/posts/qdrant-payload-index-per-key-98x-disk/6ab3877370a2e1333f041f10)

### [Natural Language as a Query Interface: qdrant-interpreter in Production Agentic RAG](https://aihive.hashnode.dev/natural-language-as-a-query-interface-qdrant-interpreter-in-production-agentic-rag?utm_source=hashnode&utm_medium=feed)

How 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, architec

25 min read · [aihive.hashnode.dev](https://aihive.hashnode.dev/natural-language-as-a-query-interface-qdrant-interpreter-in-production-agentic-rag?utm_source=hashnode&utm_medium=feed)

[Jul 12](https://hashnode.com/posts/natural-language-as-a-query-interface-qdrant-interpreter-in-production-agentic-rag/6a53495e374692fa8b0ad338)

### [The Retrieval Granularity Problem in RAG](https://aihive.hashnode.dev/the-retrieval-granularity-problem-in-rag?utm_source=hashnode&utm_medium=feed)

Why 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 i

12 min read · [aihive.hashnode.dev](https://aihive.hashnode.dev/the-retrieval-granularity-problem-in-rag?utm_source=hashnode&utm_medium=feed)

[Jul 12](https://hashnode.com/posts/the-retrieval-granularity-problem-in-rag/6a5345fc2f9c3cdbf82af7bb)