The short version
The Databricks Certified Machine Learning Professional credential is valid for two years. To keep it, you take the full exam that is live at that time; there is no short renewal quiz for this credential. If you fail, you can try again after 14 days, with no published limit on attempts.
How long the credential lasts
Databricks says a certification is valid for two years from the date it is awarded. The exam page repeats the point: recertification is required every two years to keep the credential. The digital badge follows the same clock. According to the Databricks FAQ, badges expire two years after they are awarded, and passing candidates receive them by email from a badge service, viewable on the Databricks credentials site.
The FAQ also says you will be notified in advance of your expiration date. It does not say how far in advance, so do not rely on the reminder alone. Put the date in your own calendar when the badge arrives.
How renewal works
Renewal means passing again. The FAQ puts it plainly: to recertify, you take the full current live exam. Two details follow from that.
- It is the exam of that day, not the exam you passed. Databricks updates the exam guide when the platform changes. The topics you pass today may be reorganized, trimmed, or extended in two years.
- There is no separate renewal exam for this credential. The Databricks certification page does list a renewal exam, but only for the Data Engineer Associate credential. For Machine Learning Professional, the FAQ describes only the full live exam.
Because renewal is a full attempt, it carries the same rules as a first attempt: a fee, a proctored session, and the retake wait if you do not pass. Our certification costs guide explains how to budget for exam fees without us printing amounts that change.
What happens if it lapses
Databricks does not publish a grace period. The FAQ says what you need to do to recertify, but it does not say what happens if you take the exam after the two years have ended. Based on what is published, the safe reading is this: if you let the date pass, you take the same full live exam, and you should not assume you keep any status in the meantime. If your employer or a client checks the badge date, plan to renew before it ends. If your situation depends on a grace period, ask Databricks directly rather than guessing.
Retake rules
The FAQ is clear on the basics:
| Question | What Databricks publishes |
|---|---|
| Wait between attempts | 14 days |
| Number of attempts | No limit stated; each retake needs its own 14-day wait |
| Cost | Each attempt is charged at the original exam price |
| Free retake vouchers | Not issued |
Databricks does not publish a passing score, so a failed attempt does not tell you how far you were from the line. You see unofficial results right after the exam, but the pass is only confirmed after Databricks reviews the proctor recording. Treat an unofficial pass as likely, not final.
Rescheduling is not a retake
If you simply need a different time, the FAQ sets deadlines. For online proctored exams you must reschedule more than 24 hours before the start. For test center exams the window is more than 72 hours. Missing it results in a rescheduling fee. The exam page lists both online and test center delivery, so check which applies to you when you book.
Make the 14 days count
A failed attempt is useful if you use it. Write down every topic you remember being unsure about, within an hour of leaving the exam, before the details fade. Then map each note to a topic in the domains explained page. Fourteen days is long enough for one focused round on your weakest domain and a couple of timed practice sets, but not for rebuilding a whole skill area. If your notes show gaps in several places, the study plan is a better guide than rushing back.
When the exam changes
The FAQ states that there will only ever be one version of each exam, and that the live exam guide is the authority on current topics. The Databricks exam guide for this exam says it is updated whenever the exam changes, and it asks you to check the guide two weeks before your exam date. The guide currently lists September 30, 2025 as both the publication and the last update date, with up to 10 unscored items that may appear on top of the 59 scored questions.
What does this mean for you? If you book now and the guide changes before your date, the new guide applies. Retired exams show the other side of the story. The FAQ notes that when Databricks deprecated two older exams in August 2022 because their content was outdated, credentials already issued stayed valid for the standard period. We found no statement of the same kind for this exam, and no retirement notice for it on the Databricks pages today.
A sample timeline
This is an illustration, not a promise about any individual case.
- March 3, 2027: You sit the exam. Unofficial result: pass.
- A few days later: Databricks confirms the result after reviewing the recording, and the badge arrives by email. You note the expiry as two years from the award date, around March 2029.
- February 2029: You check the exam guide for the current outline, read the changes, and book the exam for a date before the expiry.
- If you do not pass on that date: You wait 14 days, review your weak topics, and rebook. Starting early leaves room for a second attempt before the badge expires.
More guides for DATABRICKS-MLP
- Databricks ML Professional Sample Questions
- Is the Databricks ML Professional Exam Hard?
- Databricks ML Professional Prerequisites
- Is Databricks ML Professional Worth It?
- Databricks ML Professional Cheat Sheet
- Databricks ML Associate vs Professional
Next step
Read the exam format page for timing and scoring, then check where you stand with the Databricks ML Professional practice tests. For current rules, use the Databricks certification FAQ, since policies can change.
Frequently asked questions
Can I renew with a shorter exam?
Not according to the Databricks FAQ, which says you take the full current live exam. A renewal exam is listed only for a different credential.
How many times can I retake it?
Databricks states no limit, but you must wait 14 days between attempts, and each attempt is charged.
Does the exam change between passing and renewing?
It can. Databricks updates the exam guide whenever the exam changes, so check the guide before you book.