TechThiran

Most AI demos follow the same pattern. The model decides it needs a tool, the backend runs it immediately, and the result goes straight back to the model. Clean. Simple. Also a little too trusting. That works fine for prototypes. It breaks down fast in production. In real systems, you often do not want an LLM to trigger actions directly. Sometimes a tool call should pause execution instead of running right away. Maybe a human needs to approve it. Maybe another service needs t

When I began this journey, my goal was straightforward: connect an Azure Foundry agent to an Azure service using an API. However, as is often the case in software development, the details proved challenging. While Azure offers powerful tools like the Model Context Protocol (MCP) and Azure Queue Storage, neither suited my needs. MCP lacked the direct access I required, and queue-based solutions weren't ideal for the real-time integration I was pursuing. That’s when I turned t

Pagination in APIs is a fundamental concept that can significantly impact the performance and scalability of modern applications. While it’s easy to overlook, mastering pagination is essential whether you’re building a data-heavy application or preparing for a technical interview. APIs often handle massive datasets think millions or even billions of records. Without pagination, fetching data can become a nightmare. Loading an entire database in one request can lead to databa

Retrieval-Augmented Generation (RAG) frameworks have emerged as a powerful solution, allowing language models to generate responses grounded in real-world data. However, traditional retrieval methods often struggle with messy, mixed-type data or exact keyword queries. This is where Hybrid Search and Agentic RAG come into play, transforming our approach to data retrieval. In this blog, I’ll share insights from my experience working with these systems, a blend of technical e