https://ngayantha.com/blog MCP, agentic retrieval, and schema-first APIs. Practical explanations for the decisions that matter after the demo. en Fri, 02 Oct 2026 03:30:00 GMT https://ngayantha.com/blog/mcp-in-production https://ngayantha.com/blog/mcp-in-production Fri, 02 Oct 2026 03:30:00 GMT Nadeera Gayantha How Anthropic's Model Context Protocol (MCP) standardizes how LLM apps reach databases, APIs, and local tools, and how to design MCP tools with clear limits. MCP AI agents Tool design LLM security https://ngayantha.com/blog/beyond-vector-search https://ngayantha.com/blog/beyond-vector-search Thu, 27 Aug 2026 03:30:00 GMT Nadeera Gayantha Beyond naive RAG: hard filters, hybrid BM25 and dense retrieval, reciprocal rank fusion, reranking, and agentic retrieval that respects business rules. RAG Hybrid search Reranking Agentic retrieval https://ngayantha.com/blog/schema-first-api-contracts https://ngayantha.com/blog/schema-first-api-contracts Tue, 14 Jul 2026 03:30:00 GMT Nadeera Gayantha Why TypeScript type assertions are not validation, and how Zod, OpenAPI, authorization, and database constraints build schema-first, zero-trust APIs. TypeScript Zod API security Zero trust