Home/Certifications/Microsoft/Microsoft AI-103: Deploying Fine-Tuned Models Safely AI & Machine LearningUpdated October 11, 2026 · 8 min read A fine-tuned model can perform well in a notebook and still be an expensive or unreliable production choice. The training data may not reflect current user behavior, the deployment might carry hosting fees even when idle, and a new model version can change structured responses that downstream software expects. Microsoft Foundry offers several paths from a completed training job to inference, but the correct path depends on model compatibility, processing location, latency, and…

Home/Certifications/Microsoft/Microsoft AI-103: Controlling the Cost of Azure AI Apps AI & Machine LearningUpdated October 11, 2026 · 8 min read An Azure AI application’s bill is shaped by more than the advertised price of its main model. A user request can trigger retrieval, embeddings, multiple model calls, tool invocations, tracing, and retries, with each stage producing independent usage. When an agent loops unexpectedly or retrieves too much context, costs can rise without a corresponding improvement in the result. Sustainable cost control begins with measuring what it takes to complete a…

Home/Certifications/Microsoft/Microsoft AI-103: Chunking Documents for Reliable RAG AI & Machine LearningUpdated October 11, 2026 · 7 min read A RAG system can retrieve the right document and still give the model an unusable fragment. A policy exception might sit in the next paragraph, an invoice total can be separated from its table heading, and a code example may lose the explanation of its prerequisites. Chunking decides which pieces of source material are available to search and generation. It is therefore an editorial and data-engineering decision, not merely a token-limit setting….

Home/Certifications/Microsoft/Microsoft AI-103: Choosing Embeddings for Azure Search AI & Machine LearningUpdated October 11, 2026 · 7 min read Embedding models are often compared by vector dimension or public benchmark score, but those figures do not tell an enterprise search team whether its users will find the right material. A legal assistant must distinguish similar clauses, a support system must recognize technical shorthand, and a multilingual knowledge base must handle queries that use different vocabulary from its documents. Embedding selection affects all of those outcomes and changes how an Azure AI…

Home/Certifications/Microsoft/Microsoft AI-103: Choosing Azure AI Deployment Types AI & Machine LearningUpdated October 11, 2026 · 7 min read Two teams can select the same AI model and still need different Azure deployment configurations. One serves an internal assistant with uneven daytime traffic; the other processes sensitive documents under specific data-location rules and must meet a predictable throughput target. Deployment types influence capacity allocation, eligible regions, pricing, availability, and sometimes where processing occurs. The right choice is not simply the cheapest option in a portal menu. It starts with the workload’s…

Home/Certifications/Microsoft/Microsoft AI-103: Capacity Planning for Azure AI Services AI & Machine LearningUpdated October 11, 2026 · 7 min read Quota numbers look precise, but a production AI system needs more than an allowance for requests or tokens. Different tasks consume different amounts of model work, agents create unpredictable fan-out, and retrieval or tool services can become bottlenecks before inference reaches its limit. Choosing the right Azure deployment and capacity strategy requires understanding the actual workload envelope: token distribution, concurrency, latency objectives, regional constraints, and acceptable cost. Engineers should size the…

Home/Certifications/Microsoft/Microsoft AI-103: Canary Releases for Model Changes AI & Machine LearningUpdated October 11, 2026 · 8 min read A model update may pass every offline benchmark and still disappoint a small but important group of real users. New prompts, retrieval settings, or tool definitions can behave differently with messy production inputs, especially when those changes interact with unseen workflows. Canary releases let Azure AI teams expose a limited, measured slice of traffic before shifting the wider application. The objective is not simply to send ten percent of requests to a…

Home/Certifications/Microsoft/Microsoft AI-103: Designing Useful Multi-Agent Workflows AI & Machine LearningUpdated October 11, 2026 · 8 min read Splitting an application into several AI agents can make a diagram look sophisticated without improving the result. Every handoff adds latency, cost, state management, and another opportunity for tools to act under the wrong authority. Multi-agent design makes sense when distinct responsibilities genuinely need different tools, knowledge, review rules, or execution schedules. Microsoft Foundry supports several ways to build agents and orchestrate tool use, but the engineering decision begins with the workflow itself….

Home/Certifications/Microsoft/Microsoft AI-103: Blue-Green Releases for Azure AI AI & Machine LearningUpdated October 11, 2026 · 8 min read An AI deployment can be healthy by ordinary infrastructure measures and still be unsuitable for users. A changed model may interpret the same prompt differently, an embedding update may move relevant documents down the ranking, and an agent tool schema can introduce unexpected actions. Blue-green delivery gives teams a way to build and evaluate a replacement environment before routing customers to it. The technique is useful only when the two environments are…

Home/Certifications/Microsoft/Microsoft AI-103: Operating Azure AI Search for RAG AI & Machine LearningUpdated October 11, 2026 · 8 min read Adding vectors to a search index does not automatically produce trustworthy retrieval-augmented answers. The application still has to ingest documents, attach meaningful metadata, apply access controls, choose retrieval methods, and prove that relevant evidence reaches the model. Azure AI Search combines lexical and vector retrieval, filters, semantic ranking, and indexing pipelines, making it a practical retrieval layer for Microsoft Foundry applications. Those capabilities work only as well as their configuration and…

Home/Certifications/Microsoft/Microsoft AI-103: Selecting Models for Real Azure Workloads AI & Machine LearningUpdated October 11, 2026 · 8 min read The largest model in a catalog is not automatically the right one for an application. A support assistant may need fast, inexpensive grounding; a document workflow may need consistent structured extraction; and a tool-using agent may require reliable parameter selection more than elaborate prose. Microsoft Foundry makes several model families and deployment options available, but selection is constrained by region, quota, cost, latency, data governance, and observed task quality. Engineers should…

Home/Certifications/Microsoft/Microsoft AI-103: Choosing Safety Controls for Azure AI AI & Machine LearningUpdated October 11, 2026 · 8 min read Turning on a content filter is not a complete safety architecture. A customer-facing AI assistant can fail by generating harmful text, following malicious instructions embedded in a retrieved file, exposing personal information, or taking an unauthorized action. Those are distinct problems with different controls and different ways to test them. Microsoft Foundry and Azure AI Content Safety provide useful detection and evaluation capabilities, but engineers must decide where each control sits…

Home/Certifications/Microsoft/Microsoft AI-103: Identity Boundaries for Foundry Agents AI & Machine LearningUpdated October 11, 2026 · 9 min read An agent can have an excellent prompt and still become a security problem when its tools reach the wrong data. Calling a search index, updating a customer record, or sending information to another service all require authority that the model must not invent. Microsoft Foundry assigns identities differently to new-model and legacy agents, while user and downstream-tool authorization must remain distinct from the model itself. Engineers need to understand which principal makes…

ServiceNow Platform Engineering is the discipline of designing the Now Platform so data, configuration, services, integrations, automation, security, and operations remain maintainable as the enterprise grows. The platform is not only a collection of workflows. It is a shared data and execution environment whose value depends on stable models, trusted configuration data, clear ownership, reusable standards, and change practices that let many teams build without fragmenting the platform. For the Data Foundations cluster, the Configuration Management Database is one of the most important shared platform services. A trustworthy CMDB needs…

Production machine learning on AWS is the engineering of a model lifecycle that can be repeated, measured, governed, and operated. A notebook can prove that a model concept works, but a production service must also control data and feature versions, training cost, pipeline execution, artifact approval, deployment, monitoring, security, and recovery. The model is only one component of the system. Home/Amazon/Production ML on AWSAmazonPublished October 7, 2026On this pageMake reproducibility the entry requirementTreat features as production contractsOrchestrate the lifecycle instead of scripting itControl cost across training and servingDeploy the model…