CertLabProMicrosoft Azure AI Engineer Associate
225 preguntas de práctica
Última revisión: April 2026
Notas personales y enlaces de recursos para tu camino de estudio
Filtrar por Certificación
AI-102 validates the implementation skills of an Azure AI engineer: building, deploying, and operating AI solutions on Azure AI services and Azure OpenAI Service. The audience is professional developers and ML-adjacent engineers who write production code against Azure AI services rather than data scientists training models from scratch. Recent refreshes have shifted weight heavily toward generative AI, agentic solutions, and knowledge mining (RAG with Azure AI Search). Expect 40–60 questions in 100 minutes including code-completion drag-and-drops, scenario items, and case studies — heavier on SDK fluency than AI-900 or DP-100.
Largest domain at 22%. Choosing the right Azure AI service per workload, regional / model availability, capacity and TPM management, responsible-AI controls, content filters, monitoring, and cost optimization for AI workloads.
About 18%. Azure OpenAI Service (deployment, prompt engineering, function calling, structured outputs), Azure AI Foundry, model selection (GPT-4o, GPT-4 Turbo, embedding models), and content-safety integration.
About 8%. Newest area added in the 2025 refresh. Azure AI Agent Service, multi-agent orchestration, tool / function calling, state management, and safety patterns for autonomous agents.
About 12%. Azure AI Vision (image classification, object detection, OCR), Azure AI Document Intelligence (prebuilt and custom models), Azure AI Face, and Azure AI Custom Vision.
About 20%. Azure AI Language (sentiment, NER, summarization, conversational language understanding, custom question answering), Azure AI Translator, and Azure AI Speech (speech-to-text, custom voice, translation).
About 20%. Azure AI Search (formerly Cognitive Search), skillsets, indexers, vector search, hybrid search, and RAG patterns combining Azure AI Search with Azure OpenAI Service.
$115k–$160k–$220k USD annual
Range covers US-based mid-to-senior AI engineers where Azure OpenAI / Azure AI services proficiency is required. Senior GenAI engineers at FAANG / unicorns often clear $260k TC. Cert is a screening signal; production GenAI experience drives the high end.
Source: levels.fyi 2025 AI / ML engineer roles, U.S. BLS OEWS May 2024 (15-2099 ML scientists, 15-1252 software developers), Glassdoor 2025. Figures are approximate; actual compensation depends on role, region, and experience.
AI-102 demand surged through 2024–2026 as enterprises shipped Microsoft Copilot extensions, internal RAG assistants, and Azure OpenAI Service workloads into production. Recruiters use it as the de-facto signal that a candidate can credibly build on Azure OpenAI, Azure AI Search, and the broader Azure AI services portfolio. It pairs naturally with AZ-204 (Developer Associate) for full-stack GenAI roles and with DP-100 for engineers who need both implementation and model-training depth. Microsoft has invested heavily in AI-102 update cadence given the pace of GenAI evolution, so demand should remain strong.
There are no formal prerequisites. Microsoft recommends one to two years of professional development experience and prior hands-on exposure to Azure. AI-900 is a useful but not required on-ramp. AZ-204 (Azure Developer Associate) is highly complementary — many AI-102 questions assume Azure-developer-level fluency with Microsoft Entra authentication, managed identities, and Azure SDK patterns.
Proficiency in C# or Python is essentially required: code-completion drag-and-drops show real SDK snippets, with Microsoft's study material balanced between the two. The official Microsoft Learn path covers all six domains in roughly 35–45 hours; expect significant additional time in a personal Azure subscription wiring up Azure OpenAI Service deployments, RAG pipelines with Azure AI Search, and content-safety filters. The exam rewards candidates who have shipped or prototyped real GenAI applications.
AI-102 sits in the Associate tier and is broadly considered one of the more challenging associate exams given the breadth of AI services and the velocity of GenAI changes. Plan on 80–120 hours of study over 8–12 weeks with professional dev experience and prior Azure exposure; longer for engineers new to AI. The exam runs about 100 minutes with 40–60 questions in multiple-choice, multiple-response, drag-and-drop (including code-completion), hot-area, and case-study formats.
The most common stumbling block is the rapidly evolving Azure OpenAI / Azure AI Foundry surface — model deployment, capacity (TPM / PTU), and safety-filter behavior change frequently, so older study material is risky. The agentic-solution domain added in 2025 is also under-covered by third-party prep at the time of writing; lean directly on Microsoft Learn for that area.
Major refresh adding the dedicated agentic-solutions domain (Azure AI Agent Service), expanding generative-AI weight, and consolidating Cognitive Services references under Azure AI services. Microsoft has signaled faster refresh cadence given the pace of GenAI changes.
Added Azure OpenAI Service deep coverage and Azure AI Search (formerly Cognitive Search) RAG patterns; renamed services to the Azure AI services umbrella.
Initial GA, replacing the retired AI-100 exam. Original outline focused on Cognitive Services, Bot Framework, and Knowledge Mining.
AI-102 (Microsoft Azure AI Engineer Associate) is a a moderately difficult exam expecting practical hands-on experience plus solid understanding of best practices Associate-level exam. Most candidates need 80–150 hours of study spread over 6–12 weeks for associate-level exams. Most candidates who score consistently above the passing threshold on practice exams pass on their first attempt.