DP-100 Retired: What Replaces the Azure Data Scientist Certification

DP-100 was the exam behind the Microsoft Certified: Azure Data Scientist Associate. Microsoft retired it on June 1, 2026. The closest current certification is the Machine Learning Operations Engineer Associate, exam AI-300, which gives you 120 minutes. It’s a step up from DP-100, and it expects you to already have a data science background.

Is DP-100 Still Available?

No. Microsoft retired DP-100 on June 1, 2026, and the Azure Data Scientist Associate certification went with it. Microsoft’s retirement list shows “Microsoft Certified: Azure Data Scientist Associate” with a retirement date of June 1, 2026.

The certification page still loads, which is why old links and courses keep sending people to it. It carries a warning: “This certification and the renewal assessment are retired.” You can’t register for the exam or renew the credential.

If you landed here from an old study guide, you’re not alone: DP-100 still shows up in a lot of course listings.

If you already earned it, you keep it for now. Microsoft says a retired certification you earned or renewed beforehand stays on your transcript in the Active Certifications section until it expires.

What Should You Take Instead of DP-100?

The closest current option is the Microsoft Certified: Machine Learning Operations Engineer Associate, tested by exam AI-300. Microsoft’s pages we read don’t call it a direct replacement, so think of it as the nearest fit.

Here’s why it fits. Microsoft says AI-300 candidates should have “a data science background with experience in Python programming.” DP-100 was for data scientists too. The tools overlap as well, since both lean on Azure Machine Learning.

The difference is the job. DP-100 was about doing data science on Azure. AI-300 is about running it reliably, with automation, monitoring and generative AI operations added in.

If your interest is building language-model features, look at the Azure AI Apps and Agents Developer Associate, exam AI-103. Our page on the retired AI-102 certification explains it.

What Did DP-100 Cover?

The retired exam measured four areas, according to Microsoft:

• Designing and preparing a machine learning solution.
• Exploring data and running experiments.
• Training and deploying models.
• Optimizing language models for AI applications.

Microsoft described the candidate as someone with “subject matter expertise in applying data science and machine learning to implement and run machine learning workloads on Azure.” The listed tools were Azure Machine Learning, MLflow, Azure AI services including Azure AI Search, and Azure AI Foundry.

That last bullet is where the exam was already shifting. Microsoft added language-model material before retiring it, and its page noted that Azure AI Foundry had become Microsoft Foundry.

What Does the AI-300 Exam Cover?

AI-300 has five areas, and they read like a pipeline from setup to tuning.

Microsoft names the tools you should know: Azure Machine Learning, Foundry, GitHub Actions, and infrastructure as code with Bicep and the Azure CLI. It also expects “an entry-level understanding of DevOps practices.”

If you’ve only trained models in notebooks, that DevOps list is the gap to close. Putting a model into production is a different skill from fitting one.

The exam is offered in English, and the time limit is 120 minutes.

How Do You Prepare for AI-300?

Build the habit of shipping, not just training. Here’s a sensible order.

1. Refresh your Python and machine learning basics. You should be able to train, evaluate and save a simple model without looking up every step.

2. Learn the DevOps pieces. Practice a GitHub Actions workflow that runs tests, and learn what the Azure CLI and Bicep do.

3. Work in Azure Machine Learning. Train a model there, register it and deploy an endpoint. Then repeat it with a script so it runs without you clicking.

4. Try Foundry for generative AI. Deploy a model, add an evaluation and look at the monitoring output.

5. Use the practice assessment and study guide. Microsoft links the AI-300 study guide, and the practice assessment lives on AI Skills Navigator, where you sign in to use it.

Plan for months, not days, if DevOps is new to you. The exam assumes you’ve done this work, and a few evenings won’t make up for that.

What Is MLOps, and Why Did Microsoft Build an Exam Around It?

MLOps means running machine learning the way a team runs software. Models get versioned, tested, deployed through an automated pipeline and watched once they’re live. Without it, a model that worked on your laptop can quietly stop working in production.

Microsoft’s AI-300 page extends the idea with GenAIOps, which applies the same discipline to generative AI apps and agents. Microsoft groups the two under the name AI operations, or AIOps.

That’s a hint about where employers are spending. Plenty of teams can train a model. Fewer can keep one running, measure its quality and fix it when the data shifts. An exam in that area signals a skill that’s harder to find.

Should You Take AI-300 or AI-103 After DP-100?

Pick based on what you want to do all day.

Choose AI-300 if you like pipelines, automation and keeping models healthy. It matches data scientists who are drifting toward engineering. You’d lean on Azure Machine Learning, GitHub Actions and Bicep.

Choose AI-103 if you’d rather build features on top of existing models, such as chat tools, document readers and agents. It expects Python and familiarity with general AI, generative AI and Azure services, and you get 120 minutes. It doesn’t ask for the DevOps background that AI-300 does.

Some people never take either. If your work is analysis, experiments and communicating results, a stronger statistics background or a graduate degree may do more for your career than a second Microsoft exam.

How Do You Show Your Skills Now That the Exam Is Gone?

A retired credential still tells a story, but you’ll want proof that’s current. Here’s what works.

Keep the line on your resume. Write “Microsoft Certified: Azure Data Scientist Associate (DP-100), earned 20XX” with the real year. Don’t present it as active.

Build a deployed project. Train a model, deploy it behind an endpoint and write up what you monitored. That demonstrates the skills AI-300 tests, without an exam.

Use open tools too. MLflow shows up in Microsoft’s old DP-100 description, and it isn’t tied to Azure. Familiarity with it travels well between employers.

Add the degree. If you don’t have one, this is the strongest move, and it’s why the ranking lists above exist.

What If You Never Took DP-100 and Want a Data Science Career?

Start with the foundation, not an exam. Data science jobs are built on statistics, programming and data handling, and employers hire on those first. Our data scientist career guide lays out the pay, the daily work and the degree that fits.

A sensible path looks like this. Learn Python and basic statistics. Do two or three small projects with real data. Enroll in a degree, or a graduate certificate if you already hold a bachelor’s, to fill in the math. Then add a cloud credential for the platform your target employer uses.

Notice where the exam sits in that list. It comes last, after you have something to show.

The fundamentals exam, AI-901, is a gentler first credential if you want one early. Our AI-901 page explains what it covers and what happened to AI-900.

If you’re comparing a degree against a short exam, our AI certifications overview sets out how vendor exams and university certificates differ in cost, time and what employers read into each.

Does a Certification Replace a Data Science Degree?

No. A Microsoft exam shows you can use Microsoft’s tools. A degree shows you understand the statistics, linear algebra and programming underneath, and that’s what lets you choose the right model instead of the one the tool defaults to.

The DP-100 retirement makes the point. If you built your resume around that exam, you’ve now got a retired credential. A degree doesn’t expire when a vendor reshuffles its lineup.

Microsoft lists no degree requirement for AI-300 on its page, but it does ask for a data science background. That background usually comes from coursework or work experience. Our data scientist career guide covers the education employers expect.

Treat the exam as an add-on. A degree gets you into the interview, and a credential shows you’ve used the platform that employer runs.

Which Online AI Degree Lists Fit a Data Science Path?

Two of our ranking lists match this path best.

• Best Online Machine Learning Degrees is the closest fit. Programs on it go deeper into models, statistics and math, which is the foundation DP-100 and AI-300 both assume.
• Best Online Applied AI Degrees fits if you want to apply machine learning to business problems more than study theory.

If you’re starting from another field, our guide on the requirements for AI degrees shows the math and programming schools usually want, and whether you can fill the gaps.

A good test for any program: look at the course list. You want statistics, programming and machine learning courses, ideally with a project where you deploy something. That puts you in a position to take AI-300 later and understand it.

What Jobs Does a Machine Learning Certification Help With?

It helps with roles that run machine learning on Azure. Our machine learning engineer career guide describes the day-to-day work of that job, and AI-300 lines up with it closely.

The retired DP-100 pointed at data scientist roles. Those jobs focus on analysis, experiments and models, and many employers care more about your degree and portfolio than about a vendor exam. Roles for a data engineer overlap with the pipeline side of AI-300 too.

Before paying for an exam, read a few postings for the job you want. If they name Azure, Azure Machine Learning or MLOps, an AI-300 credential could stand out. If they don’t, a project that deploys a model will do more for you.

What Are the Downsides of Aiming at AI-300?

It’s a harder exam than DP-100 was. It adds DevOps and generative AI operations on top of machine learning, and Microsoft expects real experience.

It’s new. Check the dates on study material, since Microsoft updates content often.

It’s one vendor. A Microsoft credential helps most at employers that run on Azure.

Exams come and go. DP-100 retired in 2026, and the exam that follows it looks different. Build skills that outlast any exam code.

None of this means you should skip it. It means the exam is a supplement, not a plan.

How We Checked These Facts

We read Microsoft Learn’s pages for the Azure Data Scientist Associate, the Machine Learning Operations Engineer Associate and the Azure AI Apps and Agents Developer Associate on October 1, 2026. The retirement date comes from Microsoft’s retired certification exams list: “DP-100 Microsoft Certified: Azure Data Scientist Associate June 01, 2026.”

Microsoft’s pages we read showed no exam price, so confirm it with Pearson VUE before you register. Prices vary by country, and Microsoft can change names, codes and dates.