Menu
Close
RAG pipelines and multi-agent systems engineer. I approach AI through a builder's lens — interested in both the architecture and the outcome.
2024 © Edition
20:42 IST (GMT+5:30)
My collected thoughts on design, development, and everything in between.
What if instead of apps, you had agents? Instead of a file system, you had a vector store? A thought experiment about replacing the OS with an AI-native computing layer.
The gap between a RAG demo and a RAG system that handles real documents, real queries, and real users is enormous. Here is what I learned building ContextOS.
LangChain is great for prototypes. LangGraph is what you reach for when your agent needs to actually think in steps. A comparison from someone who has used both in production.
