Mihail Mateev

This session shows how to build AI-powered applications by combining ChatGPT with vector search in Azure Cosmos DB. You’ll learn how vector embeddings enable efficient similarity search over your data and how integrating Azure OpenAI models allows applications to retrieve relevant context, reason over it, and generate smarter, more accurate responses for real-world scenarios such as NLP, computer vision, and intelligent search.

This session explores how Azure Digital Twins, IoT telemetry, Microsoft Fabric, and AI agents can be combined to build fully autonomous digital twin systems. You will learn how agentic workflows continuously sense real-time data, reason over digital twin graphs with semantic context, and take automated actions

This session introduces a practical, stage-based approach to building AI agents with Azure AI Foundry. Starting from a single, task-focused agent, it demonstrates how to incrementally evolve toward an orchestrator-based architecture that coordinates specialized agents. The focus is on avoiding premature complexity, improving maintainability, and designing efficient, scalable agent systems fully aligned with the Microsoft AI stack.

SQL skills are more relevant than ever—but modern data platforms demand real-time processing, cloud scale, and event-driven intelligence. This session shows how SQL professionals can evolve existing T-SQL knowledge into scalable, real-time architectures using Microsoft Fabric. You’ll see how familiar SQL concepts extend naturally into Eventhouse, streaming analytics, and automated actions, enabling low-latency insights without abandoning proven data skills.

Vector search is becoming a core capability for AI-powered applications, but traditional SQL indexing techniques are not designed for high-dimensional similarity search. This session explains how vector search works in SQL Server, why classical indexes fail, and which practical indexing and hybrid search strategies actually improve performance. Attendees will learn how to combine relational filtering with vector similarity to build scalable, enterprise-ready AI solutions.

Next-generation digital twins require more than static models—they need real-time data, AI-driven insight, and cloud-scale architectures. This session demonstrates how to design and build operational digital twins using Microsoft Fabric, combining streaming IoT data, real-time analytics, and AI enrichment to keep twins continuously accurate, intelligent, and responsive. You’ll walk away with proven architectural patterns and practical examples for moving from dashboards to living, learning systems.

Azure OpenAI Service has introduced an image analysis feature that leverages large language models (LLMs) to comprehend the content of images. This session covers real-life cases and demos using Azure Solutions and GPT-4 Turbo with Vision and CPT-4o for enhanced predictive analysis.