LLM - 1

2026 (4)

February (2)

ai: Beyond the Prompt – Why and How to Fine-tune Your Own Models

Microsoft AI Foundry Blog

2026-02-11

Microsoft Foundry argues that enterprise LLM deployments often need behavior-level alignment (consistency, reliability, policy compliance) that prompt engineering and RAG cannot provide, positioning fine-tuning as the mechanism to change model behavior.

security: Detecting backdoored language models at scale

Microsoft Security Blog

2026-02-04

Microsoft published research on detecting backdoors in open-weight language models and described a scanner intended to identify backdoored models at scale.

January (2)

dotnet: .NET AI Essentials – The Core Building Blocks Explained

.NET

2026-01-28

The article introduces Microsoft.Extensions.AI, a unified, provider-agnostic API for building intelligent .NET applications that work with any LLM provider. It highlights a single interface that simplifies integration and includes built-in middleware, telemetry, structured outputs, and other capabilities to streamline AI development in .NET.

dotnet: Generative AI with Large Language Models in C# in 2026

.NET

2026-01-05

Practical introduction for .NET developers to using generative AI and large language models from C# in 2026, covering modern tooling, SDKs, patterns, and best practices to build safe, efficient, and maintainable AI-enabled applications on the .NET platform.

2025 (2)

November (2)

github: Enterprise bring your own key (BYOK) for GitHub Copilot is now in public preview

GitHub

2025-11-20

GitHub announced that Enterprise Bring Your Own Key (BYOK) for GitHub Copilot is now in public preview. The feature responds to enterprise demand for greater flexibility, enabling teams that already use specific LLM providers or custom models to integrate those workflows with Copilot.

azure: [In preview] Public Preview: LLM Speech in Microsoft Foundry

Azure Updates

2025-11-18

Microsoft Foundry announces a public preview of LLM Speech, an API that applies large language model capabilities to speech tasks—providing fluent, context-aware transcription and translation with expanded multilingual support and advanced prompt tuning.