Azure best practices - Architecting complex solutions with heavy loads
Speaker: Damir Dobric
Abstract:
This presentation addresses the challenges of integrating ML/AI into software projects and introduces ML.NET, Microsoft’s machine learning framework, as a solution. Various aspects of such a solution will be discussed, including proper architecture, handling heavy workloads like the ML training process, and overcoming challenges in the prediction component.
The talk provides advanced insights into the world of ML and Microsoft Azure and also imparts knowledge on handling heavy solutions in Azure, regardless of the specific topic of machine learning.
Deep Neuronal Networks vNext - Neocortex API
Speaker: Damir Dobric
Abstract:
The presentation discusses the Hierarchical Temporal Memory Cortical Learning Algorithm (HTM CLA) as a promising approach for developing true intelligent systems beyond Deep Neural Networks (DNN) and other well-known ML algorithms.
The theoretical background of the cortical algorithm and its implementation in .NET Core will be presented, including a parallel version utilizing the Actor Programming Model and Azure Service Bus. This session is aimed at architects, developers, and data scientists who want to learn more about the cortical algorithm and the latest research in the field of Computational Intelligence and Hierarchical Temporal Memory.
ML.NET - Best Practices
Speaker: Damir Dobric
Abstract:
The presentation addresses the challenges of integrating ML/AI into software projects and introduces ML.NET, Microsoft’s machine learning framework, as a solution. Various aspects of such a solution will be discussed, including the right architecture, handling heavy workloads such as the ML training process, and overcoming challenges in the prediction component. The presentation provides advanced insights into the world of ML and Microsoft Azure, and also imparts knowledge on handling heavy solutions in Azure, irrespective of the specific machine learning topic.
Machine Learning for .NET Developers
Speaker: Damir Dobric
Abstract:
ML and AI are undoubtedly becoming more important. However, integrating ML/AI capabilities into software is not an easy task. Most projects in this context are prototyped using Python. ML.NET aims to change that. In this session, you will learn how real-world projects can be implemented using the Microsoft Machine Learning .NET framework without the need to delve into the complexities of complex algorithms. ML.NET is a free, cross-platform, open-source machine learning framework for .NET developers. This session will focus on the fundamentals of the theory behind machine learning and help you get started with ML.NET. It is aimed at software architects and developers who aim to implement modern AI applications.
Modern serverless applications with Azure Functions
Speaker: Damir Dobric
Abstract:
This session highlights the benefits of serverless services in modern digital transformation and the growing popularity of cloud technology. It explains how Azure Functions can be utilized to meet various real-life requirements and operate efficiently.
The target audience for this session is software architects and developers who want to design and implement modern applications that are part of a complex distributed system across cloud and edge. Demos based on experiences and best practices from real enterprise customer projects will be presented.
How to build modern backends
Speaker: Damir Dobric
Abstract:
This session is about building advanced and modern backends using the Actor Programming Model and Azure Service Bus. It will demonstrate how a lightweight Actor Model framework based on Azure Service Bus can be utilized to develop robust, scalable, and partitioned services. Participants will learn the theory of the Actor Programming Model as well as powerful messaging patterns and experience their implementation in practice through demos.
Get started with Distributed Application Runtime (DAPR)
Speaker: Damir Dobric
Abstract:
In an era of microservices, cloud, and machine learning, the complexity of backend applications is increasing. Designing robust, reliable, scalable, and easy-to-operate systems is becoming more challenging.
This session revolves around DAPR, a new open-source runtime for distributed applications that helps in building a microservices architecture with scalability, loose coupling of services, and independent deployments made easy. Join this session to learn more about an event-driven, portable runtime environment for developing microservices for cloud and edge.
Modern backends with ‘dapr’ and ‘tye’
Speaker: Damir Dobric
Abstract:
In a time of microservices, cloud computing, and machine learning, the complexity of backend applications is increasing. It is becoming increasingly difficult to design robust, reliable, scalable, and easy-to-operate systems. This session focuses on two new open-source projects from Microsoft: the distributed application runtime Dapr, which enables the easy construction of microservice architectures with scalability, loose coupling of services, and independent deployments, and the developer tool Tye, which simplifies the development, testing, and deployment of microservices and distributed applications. Join this session to learn how to simplify microservices for the cloud and edge.
Best Practices - MongoDB in Azure CosmosDB
Speaker: Damir Dobric
Abstract:
This session presents best practices for using MongoDB in Azure CosmosDB, aiming to assist software architects and developers who are building solutions for the Microsoft Azure Cloud in leveraging the benefits of NoSQL databases in the modern world of digital transformation and cloud technology.
Advanced Backend - Actor Programming Model with dapr
Speaker: Damir Dobric
Abstract:
In this session, the Actor Programming Model is introduced, enabling software architects and developers to design complex backend systems more easily in the era of microservices, cloud, and machine learning.
Participants will learn about the Actor Programming Model and receive a practical introduction to implementing the Actor Programming Model framework based on the “dapr” framework, using numerous examples. This session is aimed at software architects and developers who want to design and implement modern applications that are part of a more complex system and are distributed across the cloud and edge.
Running .NET in containers on Azure and the Edge
Speaker: Damir Dobric
Abstract:
This session is primarily aimed at .NET developers who design, develop, and operate professional enterprise solutions.
You will learn how to package your .NET code into containers and run them locally on Azure and edge devices. Through numerous demos, you will gain insights into Docker, Azure Container Registry, and Azure Container Instances, which are core technologies for various serverless container hosts such as Kubernetes on Azure, Service Fabric, and Azure Container Apps. The cloud-native market landscape will be illuminated, and further innovations in this area will be discussed.
Modern Industrial IoT solutions with Microsoft Azure and Low Power IoT
Speaker: Damir Dobric
Abstract:
This session focuses on building solutions for various scenarios such as smart cities, logistics, supply chains, insurance, etc., using low-power and ultra-narrowband technologies. Through demos, it will demonstrate how the SigFox technology can be specifically utilized to collect telemetry data from transmitters and process it in the Microsoft Azure platform. The focus is on low-power IoT scenarios that only use small amounts of telemetry data but allow for extremely long sensor battery life. All solution architects, developers, and technical decision-makers are invited to attend.
Building modern Cloud APIs with Rate Limitation
Speaker: Damir Dobric
Abstract:
In this session, you will learn how to enhance your modern API with a rate limitation feature to restrict the number of requests based on the acquired license or to prevent DoS attacks.
Using numerous demos, we will first explain how API key authentication can be integrated into your REST API. Then, we will delve into rate limitation and demonstrate how it can be implemented using ASP.NET 7.0. This session is aimed at software architects, developers, and low-(no)-code developers.
Introduction to Azure OpenAI
Speaker: Damir Dobric
Abstract:
In this session, participants will receive a comprehensive introduction to the world of Azure OpenAI, a service that provides powerful pre-trained AI models at production scale for businesses and developers.
They will learn about various impressive technologies such as machine learning, deep learning, and natural language processing, as well as the approaches used to develop state-of-the-art AI systems. The session is aimed at architects, developers, researchers, and anyone interested in artificial intelligence.
What's New in Azure IoT?
Sprecher: Damir Dobric
Abstrakt:
In this session, participants will get an overview of the latest developments in the Azure IoT platform and learn about the new services and technologies available in Azure IoT and IoT Edge.
Through numerous demos, they will gain a comprehensive understanding of the latest advancements in Azure IoT and how they can be used to develop and deploy IoT solutions. The session is aimed at architects and developers from various industries who are interested in IoT solution development.
Azure Open AI for professional Enterprise Developers
Speaker: Damir Dobric
Abstract:
In this demo-based session, we will showcase how .NET developers can leverage Azure Open AI to build applications that go beyond traditional development methods.
We will start with an overview of the Open AI capabilities of Azure and how they can be used with any programming language, not just .NET. Then, we will dive into multiple demos that demonstrate the development of intelligent applications using Azure Cognitive Services like OpenAI.
Building intelligent applications with Large Language Models and Semantic Kernel
Speaker: Damir Dobric
Abstract:
In this demo-based session, we will show how .NET and other developers can leverage the Microsoft Semantic Kernel to harness the Azure OpenAI capabilities and build applications that cannot be implemented with traditional development methods.
We will start with an overview of Azure’s Open AI capabilities and how they can be used with any programming language, not just .NET. Then, we will delve deep into the Semantic Kernel and demonstrate how intelligent applications can be created.
Modern Serverless Applications with Azure Durable Functions, ACI and Docker
Speaker: Damir Dobric
Abstract:
In this session, you will learn how to implement and operate long-running applications using Azure Functions. You will discover how to develop and monitor long-running applications for integration scenarios.
This deep-dive session will demonstrate how to use Durable Functions to implement various complex scenarios. You will learn about orchestrations, actor models, and durable entities. Additionally, you will gain knowledge on creating and monitoring long-running processes in Docker containers running on Azure Container Registry.
Coding Workflows beyond limits of Hyperautomation – Best of Azure Durable Functions and Actor Model
Speaker: Damir Dobric
Abstract:
In today’s world of digital transformation and the growing popularity of cloud technologies, nearly every application becomes part of a more complex distributed system.
This fact increasingly requires the use of long-running workflows, state management, and system integration. In this session, you will learn how to utilize Azure Durable Functions to implement workflows using the Actor Programming Model.
ChatGPT, GPT-4 and friends, large AI models that will change our life
Speaker: Damir Dobric
Abstract:
In this talk, we will explore how large language models like ChatGPT, GPT-4, and similar models are developed and the possibilities they offer. We will discuss various applications such as text generation, translation, grammar correction, question answering, content classification, and even converting natural language to code. We will also showcase demos of OpenAI Codex and GitHub CoPilot, which can assist in writing code. Additionally, we will address the risks involved, such as potential biases. By the end of this session, participants will have a better understanding of large language models and their practical applications beyond what major providers communicate to them.
Fantastic Machine Learning models and how to find and bind them
Speaker: Damir Dobric
Abstract:
Abstract: This session showcases examples to participants on where to find and access available models for use in their own applications.
Steps for integrating and binding the models in their own code will also be demonstrated, bypassing the time-consuming process of designing and training custom machine learning models.