If you are a data scientist working with Azure Machine Learning or planning to certify in Microsoft’s data science track in 2026, the ground has shifted beneath your certification roadmap. Microsoft Azure Data Scientist Associate (DP-100) — the certification that has defined the Azure data science career path for years — is scheduled to retire on June 30, 2026. Its replacement, DP-750 (Microsoft Certified: Azure Databricks and MLOps Engineer Associate), represents a significant evolution in what Microsoft expects certified data scientists and ML engineers to know and do.
This is not a simple exam refresh. The shift from DP-100 to DP-750 reflects a real change in how organizations are building, deploying, and operating machine learning and AI solutions in production environments. DP-100 was built around Azure Machine Learning as the primary platform for data science work. DP-750 is built around the reality that modern ML engineering increasingly involves Databricks for scalable data pipelines, MLOps practices for production model management, and GenAIOps for deploying and operating generative AI solutions at enterprise scale.
For data scientists and ML engineers this creates a genuine decision point. Should you push through and earn DP-100 before it retires? Or should you invest in DP-750 and align your certification with where data science and ML engineering is actually heading?
This guide gives you the complete answer.
The Short Answer First
Take DP-100 if you are already well into your preparation, you can realistically sit the exam before June 30, 2026, and your current work is primarily centered on Azure Machine Learning for model training, experimentation, and deployment.
Take DP-750 if you are starting from scratch, your work involves Databricks, MLOps practices, or generative AI operations, and you want a certification that reflects where enterprise ML engineering is heading in 2026 and beyond.
The critical timing reality: DP-100 retires on June 30, 2026 — the same date as AI-900, AI-102, and AZ-204. If you are considering finishing DP-100, your window is tight and your preparation timeline needs to be realistic. The exam requires substantial hands-on experience that cannot be rushed.
What Is DP-100 and What Does It Actually Test
DP-100 is the Microsoft Azure Data Scientist Associate certification. It has been the primary Microsoft certification for data scientists working in Azure environments since its launch and is recognized across enterprise data science teams, consulting organizations, and technology companies as proof of practical Azure ML competency.
The certification validates that you can apply your data science and machine learning knowledge to implement and run machine learning workloads on Azure, with a primary focus on Azure Machine Learning. It covers the full data science lifecycle from data preparation through model training, evaluation, deployment, and monitoring.
The official skills measured in DP-100 cover four main domains:
Design and prepare a machine learning solution
This covers designing a data ingestion strategy for machine learning projects, designing a machine learning model development solution, creating an Azure Machine Learning workspace and required resources, managing Azure Machine Learning assets, and setting up an Azure Machine Learning development environment. It tests understanding of how to architect and configure the Azure ML environment before any training begins.
Explore data and train models
This includes exploring data for machine learning with Azure Machine Learning, training models using scripts in Azure Machine Learning, tracking model training in Azure Machine Learning, and tuning hyperparameters with Azure Machine Learning. Practical experience with Python-based machine learning workflows, experiment tracking, and automated ML is essential for this domain.
Prepare a model for deployment
This covers running pipelines in Azure Machine Learning, managing and reviewing models in Azure Machine Learning, and selecting a model deployment option in Azure Machine Learning. Candidates need to understand the full pipeline from trained model to deployment-ready artifact.
Deploy and retrain a model
This includes deploying a model to a real-time endpoint, deploying a model to a batch endpoint, and finding the best model with Automated Machine Learning. Operational deployment patterns and monitoring deployed models are significant parts of this domain.
DP-100 is considered a technically demanding certification that requires genuine hands-on experience with Azure Machine Learning, Python for data science, and machine learning concepts. Most candidates report needing 80 to 120 hours of focused preparation including significant time working directly in Azure Machine Learning environments.
What Is DP-750 and What Will It Actually Test
DP-750 is the Microsoft Certified: Azure Databricks and MLOps Engineer Associate certification. It represents a meaningful expansion and reorientation of Microsoft’s data science certification track toward the tools, practices, and operational patterns that define modern enterprise ML engineering.
The name change from “Azure Data Scientist” to “Azure Databricks and MLOps Engineer” is significant. It signals three important shifts in how Microsoft defines the certified data science and ML engineering role for 2026 and beyond.
First, Databricks becomes a first-class component of the certification scope. Databricks has become one of the most widely adopted platforms for large-scale data engineering and ML workloads in enterprise environments, and its deep integration with Azure makes it a core part of the Azure data and AI platform rather than an optional third-party tool.
Second, MLOps moves from a supporting concept to a central competency. DP-100 touched on operational concerns but was primarily a data science exam. DP-750 appears designed to validate genuine MLOps engineering skills — the practices, tools, and workflows that make machine learning systems reliable, reproducible, and maintainable in production.
Third, GenAIOps — the operational discipline of deploying and managing generative AI solutions at scale — enters the certification scope. This reflects the reality that data scientists and ML engineers in 2026 are increasingly responsible not just for traditional ML models but for generative AI deployments that require their own operational practices.
Based on Microsoft’s published direction, DP-750 is expected to cover:
Azure Databricks fundamentals and architecture
Understanding the Databricks Lakehouse architecture, Delta Lake, Unity Catalog, Databricks workflows, and how Databricks integrates with Azure Data Lake Storage, Azure Machine Learning, and other Azure data services.
Scalable data pipelines for machine learning
Building and managing data pipelines that prepare, transform, and version data for machine learning training and inference at scale using Databricks and Azure data services. This is a more engineering-heavy data preparation scope than DP-100’s Azure ML-centric approach.
Model training and experimentation at scale
Training machine learning models using Databricks ML runtime, MLflow for experiment tracking, feature engineering at scale, and distributed training patterns for large datasets and complex models.
MLOps practices and CI/CD for ML
Implementing MLOps workflows including automated training pipelines, model versioning, model registry management, continuous integration and delivery for ML systems, automated testing for ML models, and monitoring model performance and data drift in production.
Deploying and operationalizing ML models
Deploying models to real-time and batch inference endpoints, managing model serving infrastructure, implementing A/B testing and canary deployments for ML models, and maintaining deployed models across their operational lifecycle.
GenAIOps for generative AI solutions
Deploying and operating generative AI solutions including fine-tuned models, RAG implementations, and AI-powered applications. Managing the operational lifecycle of generative AI systems including monitoring, cost management, safety controls, and performance optimization.
The beta exam for DP-750 is expected in March 2026 with general availability expected in May to June 2026. As with other new 2026 certifications, the prep ecosystem will build out over time as the exam matures past its beta phase.
DP-100 vs DP-750: Key Differences at a Glance
| Area | DP-100 | DP-750 |
| Current status | Active, retires June 30, 2026 | Launching March–June 2026 |
| Credential name | Azure Data Scientist Associate | Azure Databricks and MLOps Engineer Associate |
| Primary platform | Azure Machine Learning | Azure Databricks + Azure Machine Learning |
| Primary role | Data scientist | ML engineer / MLOps engineer |
| Databricks coverage | Not covered | Central and extensive |
| MLOps depth | Surface level | Core competency |
| GenAIOps coverage | Not covered | Included |
| Data pipeline engineering | Basic | Scalable, production-grade |
| Model deployment | Included | Expanded with operational practices |
| Experiment tracking | MLflow basics | MLflow at depth |
| Prep material maturity | Excellent | Still developing |
| Exam difficulty | Challenging — requires hands-on ML experience | Expected similar or higher — more engineering focus |
| Best for | Azure ML focused data scientists certifying now | ML engineers, MLOps practitioners, Databricks users |
Who Should Take DP-100 in 2026
DP-100 remains a strong and legitimate certification choice for specific groups of data professionals in 2026. Here is exactly who those people are.
Data scientists who are already well into DP-100 preparation
If you have invested significant time studying Azure Machine Learning, working through the official learning path, completing lab exercises, and building hands-on experience in Azure ML environments, switching to DP-750 now means abandoning a substantial preparation investment. DP-100 is a respected certification with strong employer recognition. If you can finish and test before June 30, 2026, completing DP-100 is almost certainly the right move.
Professionals whose primary tool is Azure Machine Learning
If your day-to-day data science work is centered on Azure Machine Learning — running experiments in the Azure ML studio, training models using compute clusters, deploying to Azure ML endpoints, and managing models in the Azure ML model registry — DP-100’s scope maps directly to your actual job responsibilities. DP-750’s Databricks-centric framing may feel less immediately relevant if your organization does not currently use Databricks heavily.
Data scientists who need immediate certification for career progression
DP-100 is well established in job descriptions, performance review criteria, and hiring requirements across enterprise data science teams. If you need a certified credential in the next few months for a promotion, a new role, or an employer requirement, DP-100 is available right now with a mature and well-documented preparation path.
Academic researchers and students with Azure ML focus
Many academic data science programs teach machine learning using Azure Machine Learning as the cloud platform of choice. Students and researchers who have built their practical experience in Azure ML environments will find DP-100’s content highly aligned with their existing knowledge. If you can test before June 30, 2026, earning DP-100 validates that experience with a recognized credential.
Who Should Take DP-750 in 2026
For data scientists and ML engineers starting fresh or whose work is already evolving toward modern ML engineering practices, DP-750 is the more strategic investment. Here is who fits that profile best.
ML engineers and MLOps practitioners
If your role title is ML engineer, MLOps engineer, data engineer with ML responsibilities, or AI platform engineer rather than pure data scientist, DP-750’s framing aligns much more directly with your professional identity and responsibilities. DP-100 was designed for data scientists who experiment and build models. DP-750 is designed for engineers who operationalize ML systems at scale.
Data professionals already working with Databricks
If your organization uses Azure Databricks for data engineering, analytics, or ML workloads, DP-750 validates expertise in the platform you actually use every day. Databricks has become the dominant platform for large-scale data and AI workloads in many enterprises. Certifying in it alongside MLOps practices is a natural fit for data professionals in Databricks-heavy environments.
Candidates starting from scratch in mid-2026
The June 30, 2026 retirement deadline for DP-100 is tight for candidates starting from zero. DP-100 requires 80 to 120 hours of serious hands-on preparation. If you are beginning your Microsoft data science certification journey in mid-2026, building toward DP-750 from the start is almost always the smarter path than rushing through DP-100 preparation under time pressure.
Data scientists moving into generative AI operations
If your data science work is expanding to include deploying and managing generative AI solutions — fine-tuned models, RAG pipelines, AI-powered applications — DP-750’s GenAIOps content validates exactly those skills. DP-100 has no equivalent coverage because generative AI operations barely existed as a formal discipline when the exam was designed.
Professionals planning a broader ML engineering certification roadmap
If your certification goals extend beyond a single associate-level credential toward expert-level credentials in data and AI, DP-750 positions you on Microsoft’s current certification track rather than a path that is already in retirement. Building your certification stack on active, supported credentials is always a more sustainable long-term strategy.
The MLOps Revolution: Why the Shift From DP-100 to DP-750 Matters
To understand why the DP-100 to DP-750 transition is more significant than a typical exam update, it helps to understand what MLOps is and why it has become so important in enterprise data science.
DP-100 was designed in an era when the primary challenge in enterprise data science was getting models built and deployed at all. Organizations were still figuring out how to use cloud ML platforms, how to structure data science teams, and how to move from Jupyter notebooks to production-grade solutions. DP-100’s focus on Azure Machine Learning as a tool for training, evaluating, and deploying models reflected that reality.
The challenge in 2026 is different. Most mature enterprises have moved past the basic deployment challenge. The new challenge is operating ML systems reliably at scale — maintaining model quality over time, managing model versioning and rollbacks, automating retraining pipelines, monitoring for data drift and model degradation, and integrating ML workflows into software engineering CI/CD practices.
This is what MLOps addresses. And it is what DP-750 is designed to validate.
The addition of Databricks to the certification scope reflects a related shift. As ML workloads have grown in scale, many organizations have found that Azure Machine Learning alone is not sufficient for their data engineering and large-scale training needs. Databricks’ Lakehouse architecture, Delta Lake format, and distributed computing capabilities have made it the platform of choice for organizations running serious ML workloads on Azure.
DP-750 acknowledges this reality by making Databricks a first-class component of the certified ML engineer’s toolkit rather than treating it as optional advanced knowledge.
For data professionals, the practical implication is clear: the skills DP-750 validates are the skills that enterprise employers are increasingly paying a premium for. MLOps engineers and Databricks-proficient ML engineers command higher salaries and have stronger career trajectories than data scientists who only know how to train models in notebooks.
How Much DP-100 Knowledge Transfers to DP-750
This is one of the most practical questions data professionals ask when evaluating their options. The honest answer requires separating what transfers from what does not.
Knowledge that transfers well
Machine learning fundamentals — Core ML concepts including supervised and unsupervised learning, model evaluation metrics, cross-validation, feature engineering, and hyperparameter tuning are foundational knowledge that applies regardless of which platform you use.
MLflow experiment tracking — DP-100 introduces MLflow for tracking experiments in Azure Machine Learning. DP-750 uses MLflow at a deeper level within the Databricks environment. Your MLflow knowledge from DP-100 preparation carries forward directly.
Model deployment concepts — Understanding real-time versus batch inference, model endpoints, and deployment architectures is conceptual knowledge that transfers across platforms even if the specific implementation tools differ.
Azure ecosystem familiarity — General Azure knowledge including resource management, Azure storage, Azure networking basics, and Azure security concepts is background knowledge that helps in any Azure-based certification.
Knowledge that does not transfer directly
Azure Machine Learning specific operations — The specific workflows, studio interface, compute clusters, environments, and pipelines in Azure Machine Learning are largely specific to DP-100’s scope. DP-750 uses different tools for many of the same operations.
Automated ML — Azure ML’s AutoML feature is a significant part of DP-100 content. Databricks has its own AutoML capabilities but they work differently. This knowledge requires relearning rather than direct transfer.
Azure ML deployment specifics — The specific deployment patterns, endpoint types, and monitoring approaches in Azure Machine Learning differ meaningfully from the deployment patterns in Databricks and DP-750’s broader MLOps scope.
The overall picture is that conceptual ML knowledge transfers well but platform-specific operational knowledge requires significant relearning. Candidates who earned DP-100 before retirement will have a meaningful foundation for DP-750 but should not underestimate the new content they need to learn.
Timing Scenarios: What You Should Do Right Now
Scenario 1: You are actively studying DP-100 with your exam scheduled
Stay on your current path. You have a concrete goal, established materials, and a defined endpoint. Complete your preparation and test before June 30, 2026 without letting DP-750’s existence distract you. A completed DP-100 on your transcript is worth far more than an incomplete DP-750 preparation.
Scenario 2: You are studying DP-100 but have not scheduled your exam yet
Assess your preparation level honestly. If you are 65 percent or more ready, schedule your exam immediately for a date before June 30, 2026 and push through to completion. If you are less than halfway through and DP-750 better matches your role and career direction, seriously evaluate whether switching makes strategic sense. Do not switch lightly — consider your investment and your role before deciding.
Scenario 3: You have not started studying yet and it is before April 2026
You have a real choice but a tight window for DP-100. Given that DP-100 requires 80 to 120 hours of hands-on preparation, starting from zero and testing before June 30 requires an intensive and disciplined schedule. If your role is primarily Azure ML focused and you can commit to that schedule immediately, DP-100 is still achievable. If your role is more ML engineering or Databricks focused, starting directly on DP-750 may be cleaner.
Scenario 4: You have not started studying yet and it is April 2026 or later
DP-750 is almost certainly your path. The timeline for DP-100 preparation from zero is not realistic at this point without extreme time pressure. DP-750 beta is available in this timeframe and general availability follows shortly. Start building toward DP-750 from the beginning.
Scenario 5: You want both certifications
Some data professionals may choose to earn DP-100 before retirement and then pursue DP-750. This gives you a complete picture of both the traditional Azure data science path and the modern MLOps engineering path. For senior data scientists, ML architects, or data science managers this dual-credential approach demonstrates deep Microsoft data and AI platform knowledge across both generations of the certification track.
Best Preparation Strategy for DP-100
If you are targeting DP-100 before the June 30, 2026 retirement deadline, here is how to prepare effectively given the timeline pressure.
Start with the official Microsoft Learn DP-100 learning path immediately
Microsoft’s free DP-100 learning path on Microsoft Learn is comprehensive and covers all four exam domains systematically. Do not skip modules — every section contributes to building the integrated understanding the exam tests.
Set up an Azure Machine Learning workspace from day one
DP-100 cannot be adequately prepared for through reading alone. You need hands-on experience in Azure Machine Learning environments. Set up a free trial Azure account, create an Azure ML workspace, and practice every concept you study by implementing it directly in the platform.
Focus on pipeline development and automation
Many DP-100 candidates underestimate the pipeline content. Azure ML pipelines — creating them, running them, scheduling them, and troubleshooting them — are a significant part of the exam. Practice building multi-step pipelines in both the designer and through code.
Master the MLflow tracking workflow
MLflow experiment tracking appears throughout DP-100 and is tested at a practical level. Understand how to log parameters, metrics, and artifacts. Practice comparing runs in the Azure ML studio and selecting the best model based on tracked metrics.
Practice real-time and batch endpoint deployment
Deployment is a critical domain. Make sure you can deploy a model to both a managed online endpoint and a batch endpoint, understand the configuration options for each, and know how to test and monitor deployed endpoints.
Validate your readiness before booking your exam
Given the time pressure, knowing your genuine readiness level before committing to an exam date is essential. Use our DP-100 exam preparation materials to assess where you stand and identify the gaps you need to close before exam day.
Best Preparation Strategy for DP-750
Since DP-750 is a new exam with beta starting in March 2026 and general availability expected in May to June 2026, here is how to prepare strategically.
Build strong Databricks fundamentals as your foundation
Databricks has its own free learning platform — Databricks Academy — with extensive free content covering the Databricks Lakehouse platform, Delta Lake, Databricks ML, and MLflow. Working through the relevant Databricks learning paths now gives you a strong foundation before official DP-750 materials are available.
Learn MLflow deeply rather than superficially
MLflow is the common thread across both Databricks ML and Azure Machine Learning. Understanding MLflow experiments, runs, models, model registry, and deployment at a deeper level than DP-100 requires will pay dividends throughout DP-750 preparation.
Study Delta Lake and the Databricks Lakehouse architecture
Delta Lake is the storage layer underlying the Databricks Lakehouse and is central to how Databricks handles data for ML workloads. Understanding ACID transactions in Delta Lake, time travel, schema enforcement, and data versioning is important foundational knowledge for DP-750.
Understand MLOps principles and CI/CD for ML
Read Microsoft’s and Databricks’ documentation on MLOps practices. Understand the end-to-end MLOps workflow from data versioning through model training, testing, deployment, and monitoring. Familiarize yourself with how Azure DevOps and GitHub Actions integrate with ML workflows.
Study GenAIOps concepts and generative AI deployment patterns
Understanding how to deploy and operate generative AI solutions — fine-tuned models, RAG pipelines, LLM-based applications — at the operational level is expected to be part of DP-750. Microsoft’s documentation on responsible AI deployment and Azure AI operations is a good starting resource.
Follow official DP-750 materials when they become available
When Microsoft publishes the official DP-750 learning path and exam blueprint on Microsoft Learn, make those your primary reference. Third-party materials will follow but always anchor your preparation in official Microsoft content.
DP-100 and DP-750 in the Broader 2026 Microsoft Data and AI Landscape
The DP-100 to DP-750 transition does not happen in isolation. It is part of Microsoft’s coordinated 2026 certification overhaul that is reshaping credentials across every technical track as we covered in our complete guide to Microsoft certifications retiring in 2026.
For data professionals the broader context is important. The developers on your team are navigating the AZ-204 to AI-200 transition. The AI engineers are deciding between AI-102 and AI-103. The security engineers are evaluating the AZ-500 to SC-500 shift. Everyone is being asked to grow their AI and modern cloud competency alongside their existing domain expertise.
For data scientists specifically, DP-750’s emphasis on MLOps and GenAIOps reflects a convergence that has been building for several years: the distinction between data scientist and ML engineer is blurring, and the professionals who can bridge both — who can build models and operate them reliably in production — are the most valuable members of modern data and AI teams.
DP-750 is Microsoft’s certification recognition of that convergence. Earning it positions you not just as someone who can train models but as someone who can build and maintain the systems that make ML reliable in enterprise environments.
Final Verdict: DP-100 or DP-750 in 2026
The right answer depends on your current preparation level, your role, and your career direction.
Choose DP-100 if:
- You are already well into preparation and can test before June 30, 2026
- Your primary work platform is Azure Machine Learning rather than Databricks
- You need an immediately recognized data science credential for career progression now
- You prefer a proven exam with a mature and complete preparation ecosystem
Choose DP-750 if:
- You are starting from scratch and cannot realistically prepare and test before June 30, 2026
- Your work involves Databricks, large-scale data pipelines, or production ML operations
- You want a certification that reflects where enterprise ML engineering is heading
- You are building toward a broader ML engineering or AI platform career path
The honest bottom line: DP-100 remains a valuable credential if you can finish it in time. DP-750 is where Microsoft’s data science and ML engineering certification track is going. The skills DP-750 validates — Databricks, MLOps, GenAIOps — are increasingly what enterprise employers need and reward.
If you are close to ready for DP-100, earn it. If you are just starting out, build toward DP-750. The investment in either direction is worthwhile as long as you make the decision deliberately and follow through completely.
FAQs
Is DP-100 retiring in 2026?
Yes. Microsoft has confirmed that the Azure Data Scientist Associate certification and its associated DP-100 exam will retire on June 30, 2026. After that date you cannot take the exam or earn the certification for the first time.
What is replacing DP-100?
DP-100 is being replaced by DP-750, the Microsoft Certified: Azure Databricks and MLOps Engineer Associate certification. DP-750 focuses on Azure Databricks, scalable ML pipelines, MLOps practices, and GenAIOps for deploying and operating generative AI solutions.
Will my DP-100 automatically become DP-750 after retirement?
No. Retired certifications are not converted to their replacements. DP-750 must be earned independently by completing and passing the DP-750 exam. Your existing DP-100 credential remains on your transcript until it naturally expires.
Is DP-750 harder than DP-100?
The full DP-750 exam blueprint is not yet published so a definitive comparison is not possible. Based on Microsoft’s announced scope, DP-750 covers MLOps engineering practices, Databricks platform depth, and GenAIOps in addition to core ML concepts. Candidates who are strong in practical ML engineering will likely find DP-750 well aligned with their experience. Candidates who are primarily theoretical data scientists may find the engineering emphasis more challenging.
Do I need Databricks experience to pass DP-750?
Based on Microsoft’s announced direction, Databricks is a central component of DP-750’s scope. Candidates without prior Databricks experience will need to invest meaningfully in learning the platform before attempting the exam. Databricks Academy offers free learning resources that can help build this foundation.
When will DP-750 training and exam materials be available?
Based on Microsoft’s announcements, DP-750 beta is expected in March 2026 with general availability expected in May to June 2026. Official Microsoft Learn training materials are expected to become available alongside the beta exam.
Is DP-100 still recognized by employers in 2026?
Yes. DP-100 remains a recognized and respected data science certification in enterprise hiring. Its employer recognition will remain strong through its retirement date and for some time afterward. However, as DP-750 establishes itself over the next 12 to 18 months, employers will increasingly look for it as the current-generation credential.
Can I take both DP-100 and DP-750?
Yes. Some data professionals may choose to earn DP-100 before retirement and then pursue DP-750. This dual-certification approach demonstrates both traditional Azure ML data science expertise and modern MLOps engineering competency. For senior professionals or those in consulting roles it is a legitimate career investment.
Where can I find DP-100 study materials right now?
Microsoft Learn has a complete free learning path for DP-100 covering all four exam domains. You can also use our DP-100 exam preparation materials to validate your readiness before your exam date.
How does DP-100 retirement fit into Microsoft’s broader 2026 changes?
DP-100 is one of seven major Microsoft certifications retiring in 2026. Our complete Microsoft certifications retiring in 2026 guide covers every retirement, replacement exam, and transition strategy across all affected tracks.
