SRSwati Rawatinagentic-ai-toadster.hashnode.dev·1d ago · 21 min readThe Layer Between Retrieval and Generation That Most RAG Systems Never BuildYour retrieval recall sits around ninety percent. The reranker is tuned. The embedding model is appropriate for your domain. You spot checked twenty queries and the correct chunks came back every time00
SRSwati Rawatinagentic-ai-toadster.hashnode.dev·2d ago · 13 min readBuilding AI Applications That Know Things and Do Things: The Case for RAG Plus MCPThere is a question that comes up early in almost every serious AI application project. "Should we use RAG or MCP?" It is the wrong question. RAG and MCP do not solve the same problem. They do not com00
SRSwati Rawatinagentic-ai-toadster.hashnode.dev·5d ago · 10 min readBuilding AI Agents That Know When to Stop and Ask a HumanThere is a moment in almost every AI agent project where someone says: "Can we just let it handle this automatically?" Sometimes the answer is yes. A lot of the time, the honest answer is: it depends 11M
SRSwati Rawatinagentic-ai-toadster.hashnode.dev·Sep 10 · 4 min readHow AI Companies in Noida Are Handling Data Privacy Under India's DPDP ActWhen India passed the Digital Personal Data Protection (DPDP) Act, it permanently altered how software companies process user information. Before DPDP, many teams treated customer data casually. Promp00
SRSwati Rawatinagentic-ai-toadster.hashnode.dev·Sep 3 · 7 min readThe Real Difference Between an AI Feature and an AI Product, and Why Most Companies Build the Former When They Need the LatterA company adds a chatbot to its existing customer support page and starts describing itself, in pitch decks and marketing copy, as an "AI-powered" business. Somewhere else, a different company builds 00
