Frameworks AI
| Framework/Platform | Description | Last Update |
|---|---|---|
| Azure OpenAI Service | Provides API access to OpenAI's advanced models like GPT-4, Codex, and DALL·E, enabling tasks such as text generation, code completion, and image creation. Hosted on Azure for scalability and security. | January 2025 (reuters.com) |
| Azure Machine Learning (Azure ML) | A comprehensive platform for building, training, and deploying custom machine learning models. Supports the entire ML lifecycle with integration into Azure services for streamlined data processing and analysis. | Ongoing updates |
| Copilot Studio | A low-code/no-code tool for creating AI-powered applications within Microsoft products like Power Apps, Teams, and Office applications. Simplifies AI development, allowing business users to build intelligent applications with minimal coding. | September 2024 (axios.com) |
| ML.NET | An open-source machine learning library for .NET developers, enabling the creation of custom ML solutions using C# or F#. Supports tasks like classification, regression, and recommendation systems, with cross-platform support. | November 2024 (multiplatform.ai) |
| DeepSpeed | An open-source deep learning optimization library designed to reduce computing power and memory usage when training large distributed models. Features the Zero Redundancy Optimizer (ZeRO) for efficient training of models with trillions of parameters. | January 2025 (theverge.com) |
| ONNX (Open Neural Network Exchange) | An open-source format for representing machine learning models, promoting interoperability between different frameworks. Allows models trained in various frameworks to be transferred and run across diverse environments. | Ongoing updates |
| Azure Cognitive Search | A cloud-based service providing indexing and querying capabilities for data uploaded to Microsoft servers. Offers AI-powered enrichments to extract text from images and other unstructured data sources, enhancing searchability. | Ongoing updates |
| Microsoft Cognitive Toolkit (CNTK) | A deprecated deep learning framework developed by Microsoft Research, previously used for describing neural networks as computational steps via a directed graph. | Deprecated |
| AutoGen | A multi-agent AI framework that enables large language models (LLMs) to collaborate and automate workflows, facilitating complex problem-solving through AI agent cooperation. | November 2024 (techcommunity.microsoft.com) |
| Semantic Kernel | An AI orchestration framework that integrates LLMs into applications and workflows. Supports prompt chaining, memory management, and AI plugin development, working with languages like Python and C#, and services like Azure OpenAI. | October 2024 (infoq.com) |
| AI Foundry | A Microsoft initiative aimed at building custom AI models for businesses, focusing on fine-tuning LLMs and embedding them into enterprise solutions to democratize AI development within organizations. | January 2025 (theverge.com) |
Semantic Kernel is great if...
✅ You want to integrate AI (LLMs) into apps and workflows easily.
✅ You need prompt chaining, memory, and AI plugins out of the box.
✅ You’re working with C# (.NET) or Python.
✅ You're using Microsoft tools (Azure OpenAI, Copilot, etc.).
Content maintained by ITConnector b.v.
