Browse All Machine Learning Content (211)

The Microsoft Fabric Blog introduces Capacity Scheduler (preview) for Eventhouse, explaining how teams can schedule minimum capacity across the week to better match predictable real-time analytics patterns while keeping autoscale enabled.

Microsoft Fabric Blog explains how Fabric’s “Analyze data with” entry points are being unified across Lakehouse, Data Warehouse, and Eventhouse (Preview), embedding Eventhouse Endpoint alongside SQL Endpoint and Notebooks to make analysis more discoverable and consistent.

Microsoft Fabric Blog introduces (in preview) how Eventhouse in Microsoft Fabric unifies real-time analytics with anomaly detection, data agent integration, SQL endpoints, and notebooks—aiming to move teams from live event signals to investigation and action without duplicating data or stitching multiple tools.

Microsoft Fabric Blog announces that Fabric Data Warehouse now supports running certain ALTER TABLE operations inside explicit transactions, enabling atomic schema changes with automatic rollback for safer deployments.

Microsoft Fabric Blog announces the general availability of OneLake File Explorer, a Windows File Explorer integration that lets teams drag and drop local files into OneLake so they can be used immediately in Microsoft Fabric pipelines, notebooks, semantic models, and downstream analytics workflows.

Microsoft Fabric Blog announces (preview) full Azure SQL Database collation support for SQL databases in Microsoft Fabric, explaining what collations are and how to set them during database creation via the Fabric REST API and other deployment methods.

Welcome back to this week's roundup. The main thread is that agents are showing up in more places, and teams are getting clearer ways to control how those agents run. Updates across Copilot (IDE, CLI, cloud agent, and mobile) focused on practical autonomy controls, offline/BYOK routing, cross-model review checkpoints, and security remediation loops that end in reviewable pull requests. In parallel, MCP and Azure AI Foundry updates continued to reinforce "run it like software" basics: deployable tool surfaces with real auth, consistent runtimes across cloud and local, and clearer observability and identity boundaries for day-two operations.

Gapandey lays out a practical, end-to-end MLOps template on Azure: train a scikit-learn model from data in Azure Blob Storage, package it as a self-contained pickle bundle, register it in an Azure ML Registry with auto-versioning, and deploy it to an Azure ML Managed Online Endpoint via an Azure DevOps multi-stage pipeline.

Microsoft Fabric Blog (with coauthor Arindam Chatterjee) summarizes Q1 2026 updates for Fabric Eventstreams, covering new connectors (DeltaFlow, MQTT v3, Anomaly Detection), tighter Spark Structured Streaming/Notebook integration, and enterprise networking and security features like private network ingestion and Key Vault-backed custom CA + mTLS.

Microsoft Fabric Blog announces a preview feature that lets you associate a user, service principal, or managed identity with Fabric items (currently Lakehouses and Eventstreams) so those items no longer depend on the original owner’s credentials.

Microsoft Fabric Blog announces the general availability of Shortcut transformations, a Fabric/OneLake feature that turns files (CSV, Parquet, JSON) referenced via shortcuts into continuously synchronized Delta tables—without building ETL pipelines or writing code.

Microsoft Fabric Blog announces a preview feature for Fabric Eventstream connectors that adds support for custom Certificate Authorities and mutual TLS, using Azure Key Vault to store and rotate certificates for Kafka-based streaming sources.

AnjaliSadhukhan argues that AI agents fail on enterprise questions mainly due to fragmented data and missing semantics, and outlines how Microsoft Fabric (OneLake, semantic models, Data Agents) and Azure AI Foundry can work together to provide governed, agent-ready access to business data.

Microsoft Fabric Blog introduces the Migration Assistant (Preview) for SQL database in Microsoft Fabric, outlining a wizard-driven process to migrate from SQL Server or Azure SQL Database using DACPAC schema import, compatibility checks, and Fabric Copy Jobs—positioning operational data for OneLake-based analytics and AI workloads.

Microsoft Fabric Blog announces general availability of ANY_VALUE() in Fabric Data Warehouse, explaining how this T-SQL aggregate helps keep GROUP BY clauses focused on the true grouping key while still returning functionally dependent descriptive columns like city, state, and country.

Microsoft Fabric Blog introduces (preview) built-in AI functions in Microsoft Fabric Data Warehouse to extract, classify, analyze sentiment, and transform unstructured text directly in T-SQL, including a prompt-based function for custom processing.

This week’s roundup has a consistent theme: agents are becoming normal participants in day-to-day workflows, and platforms are adding the guardrails that make that workable at scale. Copilot’s cloud agent added branch-first work, plan-first flows, and deep research, plus enterprise controls such as runner placement, firewall policy, and verified commit signing. CLI and SDK updates also moved toward multi-agent orchestration and reusable runtimes. Across Azure and Fabric, the same pattern shows up: more standardized orchestration and stronger day-2 operations, alongside security guidance that focuses on practical risk-reduction points like dependency installs, CI configuration, and admission-time enforcement.

Microsoft Fabric Blog announces a preview ADO.NET driver that lets .NET apps connect to Microsoft Fabric Data Engineering Spark workloads via Livy APIs, using familiar ADO.NET abstractions plus Entra ID authentication, pooling, session reuse, and Spark SQL support.

Microsoft Fabric Blog explains a Microsoft Fabric (Preview) update that lets you enable workspace-level customer-managed keys (CMK) even when the underlying Fabric capacity is configured for Bring Your Own Key (BYOK), simplifying enterprise encryption and compliance setups while using Azure Key Vault or Managed HSM.

ShivaniThadiyan explains how Azure SQL Managed Instance is evolving from a SQL Server-compatible PaaS into an AI-enabled platform, covering built-in operational intelligence, vector search, in-database Python/R machine learning, and Copilot-assisted diagnostics with security and governance considerations.