PS HarriJaakkonen :~/Blog/Posts> cat ./machine-learning-operations-engineer-ai-300-study-guide.html

AI-300 Study Guide: Machine Learning Operations Engineer Associate

AI-300 Machine Learning Operations Engineer Associate Study Guide

New certification: AI-300 replaces DP-100 and broadens the scope well beyond data science. It covers both classical MLOps (Azure Machine Learning) and GenAIOps (Microsoft Foundry, generative AI evaluation, fine-tuning). Beta exam opened April 2026, goes live May 2026.

Overview

The AI-300 exam is aimed at engineers responsible for taking AI and ML systems to production and keeping them running reliably. This includes infrastructure-as-code for ML workspaces, CI/CD pipelines for model retraining, deploying and managing large language models on Foundry, and implementing observability for both classical models and generative AI applications.

Exam Details

Detail Information
Exam code AI-300
Certification Machine Learning Operations Engineer Associate
Level Associate
Passing score 700 / 1000
Duration 120 minutes
Cost $165 USD (varies by region)
Beta availability April 2026
Exam goes live May 2026
Predecessor DP-100 (retires June 1, 2026)

Exam Objectives

The AI-300 exam is organized into five skill domains. Together, they cover the full lifecycle of machine learning and generative AI operations — from infrastructure and pipeline design through model lifecycle management, deployment, quality assurance, and performance optimization.

Design and Implement MLOps Infrastructure (15–20%)

Create and manage machine learning environments at scale. Includes:

  • Building Azure Machine Learning workspaces with Bicep or Terraform IaC
  • Configuring datastores, compute targets, and data assets for reproducible training
  • Implementing secure identity with managed identities and Azure RBAC role assignments
  • Enforcing network isolation with private endpoints, service endpoints, and firewall rules
  • Automating workspace provisioning and project workflows with GitHub Actions and Git integration
  • Setting up registries for sharing trained models, environments, and reusable components across workspaces

Implement Machine Learning Model Lifecycle and Operations (25–30%)

Manage the full model lifecycle from training through production. Covers:

  • Building training pipelines with components, jobs, hyperparameter tuning, and distributed training
  • Configuring MLflow experiment tracking, model registry versioning, and lineage tracking
  • Deploying models to managed online endpoints for real-time inference and batch endpoints for asynchronous scoring
  • Implementing progressive rollout strategies: blue/green, canary, A/B testing, and traffic management
  • Monitoring model performance, data drift, and operational health using responsible AI tooling
  • Configuring automated alerts, retraining triggers, and continuous model refresh workflows
  • Managing endpoint scaling, quotas, and cost optimization across deployments

Design and Implement GenAIOps Infrastructure (20–25%)

Build production-grade generative AI systems in Microsoft Foundry. Includes:

  • Provisioning Foundry Hubs and Projects using Bicep, Terraform, or Azure CLI
  • Securing Foundry resources with managed identities, Azure RBAC, private endpoints, and private networking
  • Deploying foundation models (GPT-4o, o3, Phi, Llama, Mistral) using serverless API endpoints, managed compute, and Provisioned Throughput Units (PTU)
  • Implementing model versioning, production deployment strategies, quota planning, and rate-limit management
  • Designing and versioning prompts in Git repositories with variant tracking and A/B testing
  • Configuring API Management (APIM) gateways for semantic caching, token limits, and policy enforcement

Implement Generative AI Quality Assurance and Observability (10–15%)

Ensure GenAI applications meet quality and safety standards. Covers:

  • Creating comprehensive test datasets and evaluation mappings for models, applications, and agents
  • Configuring evaluation runs with built-in evaluators: groundedness, relevance, coherence, fluency
  • Implementing custom evaluators using Python and LLM-as-judge patterns
  • Building quality gates into development and release CI/CD pipelines for automated validation
  • Setting up continuous monitoring for generative AI applications and agents in production
  • Tracking latency, throughput, token consumption, resource usage, and total cost of ownership
  • Configuring detailed logging, OpenTelemetry tracing, and debugging for production troubleshooting

Optimize Generative AI Systems and Model Performance (10–15%)

Improve RAG pipeline accuracy and inference efficiency. Includes:

  • Optimizing Retrieval Augmented Generation with chunk sizes, similarity thresholds, reranking, and query expansion
  • Selecting, evaluating, and fine-tuning embedding models for domain-specific retrieval quality
  • Implementing hybrid search combining semantic, vector, and keyword retrieval techniques
  • Designing and executing advanced fine-tuning workflows using Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Reinforcement Learning from Human Feedback (RLHF)
  • Creating and managing synthetic training data for fine-tuning and domain customization
  • Moving fine-tuned models from development to production with performance monitoring
  • Optimizing inference cost and latency through batching, semantic caching, model-size selection, and deployment-tier optimization

Skills Measured

Design and Implement MLOps Infrastructure (15–20%)

  • Create and manage Azure Machine Learning workspaces, datastores, and compute targets
  • Create and manage data assets, environments, and reusable components
  • Share assets across workspaces by using Azure Machine Learning registries
  • Configure identity and access management with managed identities and Azure RBAC
  • Deploy Azure Machine Learning workspaces and resources by using Bicep and Azure CLI
  • Automate provisioning and project workflows with GitHub Actions and Git integration
  • Restrict network access to Machine Learning workspaces with private networking controls

Implement Machine Learning Model Lifecycle and Operations (25–30%)

  • Configure MLflow experiment tracking and compare model performance across jobs
  • Use notebooks, automated machine learning, training scripts, hyperparameter tuning, and distributed training
  • Build training pipelines with Azure Machine Learning jobs, components, and reusable assets
  • Register, version, package, and deploy models by using MLflow and Azure Machine Learning registries
  • Deploy models to managed online endpoints for real-time inference and batch endpoints for asynchronous scoring
  • Implement progressive rollout, blue/green, canary, rollback, and endpoint traffic-management strategies
  • Evaluate model quality with Responsible AI tooling and monitor deployed models for drift, performance degradation, and operational health
  • Configure alerts, retraining triggers, and automated model refresh workflows

Design and Implement GenAIOps Infrastructure (20–25%)

  • Create and manage Microsoft Foundry resources, projects, and project environments using IaC
  • Configure managed identities, Azure RBAC, private endpoints, and private networking for Foundry workloads
  • Deploy and manage foundation models (GPT-4o, Phi, Llama) for production workloads
  • Select model deployment approaches across serverless API endpoints, managed compute, and provisioned throughput units (PTU)
  • Implement model versioning, production deployment strategies, quota management, and rate-limit planning
  • Design, develop, compare, and version prompts and prompt variants in Git repositories

Implement Generative AI Quality Assurance and Observability (10–15%)

  • Create test datasets and data mappings for comprehensive model, app, and agent evaluation
  • Configure evaluation runs in Microsoft Foundry using built-in and custom evaluators
  • Assess groundedness, relevance, coherence, fluency, safety, and harmful-content risk
  • Automate evaluation workflows as quality gates in development and release pipelines
  • Use continuous monitoring in Foundry for generative AI applications and agents
  • Monitor latency, throughput, response time, token consumption, resource usage, and cost
  • Configure detailed logging, tracing, and debugging capabilities for production troubleshooting

Optimize Generative AI Systems and Model Performance (10–15%)

  • Optimize RAG pipeline accuracy with chunk sizes, similarity thresholds, retrieval strategies, and reranking
  • Select, evaluate, and fine-tune embedding models for domain-specific retrieval quality
  • Implement hybrid search that combines semantic, vector, and keyword retrieval
  • Evaluate RAG quality with relevance metrics, A/B tests, and retrieval diagnostics
  • Design and implement advanced fine-tuning and model customization workflows
  • Create and manage synthetic data for fine-tuning
  • Move fine-tuned models from development to production and monitor their performance
  • Optimize inference cost and latency with batching, caching, model-size selection, and deployment-tier choices

Exam Domain Weights

Domain Weight Study Part Focus Area
1 — Design and Implement MLOps Infrastructure 15–20% Part 1 Workspace IaC, RBAC, network isolation, compute, data assets, MLflow, CI/CD
2 — Implement ML Model Lifecycle and Operations 25–30% Part 2 Pipeline components, registry, online/batch endpoints, monitoring, deployment strategies
3 — Design and Implement GenAIOps Infrastructure 20–25% Part 3 Foundry Hub/Project IaC, PTU vs serverless, prompt versioning, APIM gateway
4 — Implement GenAI Quality Assurance and Observability 10–15% Part 4 Built-in evaluators, custom evaluators, CI/CD quality gates, OpenTelemetry tracing
5 — Optimize Generative AI Systems and Model Performance 10–15% Part 5 RAG optimisation, fine-tuning, Batch API, semantic caching, model routing, streaming

Study priority: Domain 2 (ML Lifecycle) carries the most weight at 25–30%. Allocate study time proportionally — Domains 1 and 2 together account for 40–50% of the exam. Domain 3 (GenAIOps) has grown significantly to 20–25%, so don't underweight it. Domains 4 and 5 together are 20–30% and share many conceptual overlaps around evaluation and optimisation.

MLOps vs GenAIOps — Side by Side

AI-300: MLOps vs GenAIOps MLOps — Azure Machine Learning • IaC workspace provisioning (Bicep) • Pipeline components and jobs • MLflow tracking and model registry • Online and batch endpoints • Blue/green and canary deployments • Data drift monitoring and alerts • Automated retraining triggers • Hyperparameter sweep jobs • CI/CD with GitHub Actions / ADO ~40–50% of exam weight GenAIOps — Azure AI Foundry • Foundry Hub/Project IaC provisioning • PTU vs serverless model deployment • Prompt versioning and source control • Built-in evaluators (groundedness, safety) • Custom evaluator implementation • OpenTelemetry distributed tracing • RAG pipeline optimisation • Fine-tuning and distillation • Quota, rate limits, APIM gateway ~40–55% of exam weight Domains 1 and 2 (MLOps) account for ~40–50% of exam questions
AI-300 splits coverage between classical MLOps (Azure ML) and GenAIOps (Azure AI Foundry).

DP-100 vs AI-300 — Key Differences

Area DP-100 AI-300
Primary role Data Scientist (build and train models) MLOps Engineer (operate models in production)
IaC / CI/CD Not covered Core domain — Bicep, GitHub Actions
Generative AI Not covered Full GenAIOps domain on Foundry
Monitoring Basic data drift detection Full observability — tracing, dashboards, alerts
Model training Heavy focus (AutoML, sweep jobs) Pipeline orchestration, not hands-on training
Evaluation MLflow metrics on classical models Both classical metrics and gen AI evaluators
Fine-tuning Not covered Foundation model fine-tuning and distillation

Key Tools and Technologies

  • Azure ML Python SDK v2 — primary SDK for workspace operations, pipelines, and endpoints
  • MLflow — experiment tracking, model registry, model signatures
  • Azure AI Foundry SDK — hub/project management, model deployments, evaluations
  • GitHub Actions / Azure DevOps — CI/CD pipelines for model retraining and deployment
  • Bicep / Terraform — IaC for workspace and Foundry infrastructure
  • OpenTelemetry — distributed tracing standard used by Azure for AI observability
  • Azure Monitor / Application Insights — monitoring dashboards and alert rules

Recommended Study Path

  1. Azure ML Pipelines overview — pipelines, components, scheduling
  2. Deploy models to online endpoints — blue/green, traffic splitting
  3. MLOps and GenAIOps for AI workloads on Azure — Well-Architected Framework guidance
  4. Evaluate generative AI apps in Foundry — evaluators, metrics
  5. Official AI-300 study guide — Microsoft Learn