ML Observability

Learn how to leverage popular tools to implement ML observability for critical models to detect and explain why the performance of your production models degrades over time due to causes like model drift.

Alexey Grigorev image

Meet your instructor Alexey Grigorev

Alexey Gregoriev is the founder of DataTalks.Club (a community of 30K+ data enthusiasts), Principal Data Scientist at OLX and author of several books such as the Machine Learning Bookcamp.

Implementing ML Observability

To build ML at scale, companies need to operate multiple ML solutions in parallel. One duplicate data set or stale model can cause unintended (but severe) consequences that monitoring alone can’t catch or prevent.

Similar to how data observability helps answer the why, ML observability enables model owners to do root cause analysis of model failures. Monitoring alerts you when the performance of your model is degrading. ML observability helps you get to the bottom of why. Figuring out what needs to be done once an issue arises is the bigger, harder problem.

We’ll cover:

  • Observability and monitoring: tracking data and model quality in production
  • Orchestrating data and model quality checks
  • Designing test suites and checks for your ML models
  • Configuring alerts


  • Evidently — for model and data quality analysis
  • Prefect — for workflow orchestration and automation
  • AWS for cloud resources

By the end of this course, you’ll be able to design and implement an end-to-end ML and data observability pipeline.

A learning approach that aligns with your company values.


Bite-sized daily lessons that you can easily fit into your schedule. Each day, we release new lessons no longer than 15 minutes. Our lessons are carefully curated to ensure that they're both engaging and informative, allowing you to learn something new every day, and at your own pace.


Collaborate with other engineers from around the world, providing you with a unique opportunity to learn from others and build your professional network.


Our live learning sessions are designed to be interactive and engaging, giving you the opportunity to ask questions and interact with subject-matter experts.


Learn by solving real-world problems. Our courses are designed to get rid of the fluff and provide you with the most relevant information to help you apply your learning.

Trusted by teams from global companies

Frequently asked questions

Are all sessions live?

Yes, all sessions during the cohort will be live with the instructor. However, we will record each session and make them available for everyone in the cohort.

What is the time commitment?

Our courses typically have 2-4 modules, with each module lasting for approximately 2h per week which you can block out on your calendar during your work day. You also get some take home projects that you can complete at your own pace.

Do I earn a certificate for this course?

Of course! Once you’ve completed the course modules, you will get a certificate of completion that you can showcase to the world.

What is included in a LearnCrunch membership?

With your LearnCrunch yearly membership, you get access to our live instructor-led cohorts, our catalog of self-guided courses, unlimited real-world projects to learn from and master new skills, exclusive live events for members and a vetted global community of experts and peers.

How much does a LearnCrunch membership cost?

An individual LearnCrunch membership costs $1,000 USD per year. If your company is interested in purchasing multiple seats, please contact

Can I expense this course?

Yes. Most LearnCrunch members have expensed this course through their Learning & Development budget, similar to how you expense conferences. You can use this email template to request expense approval from your manager.

I have more questions. Get in touch with us!

If you have more questions, email us at

Book a call with us

Victor Chima

Co-Founder at LearnCrunch

Connect with us to learn how we can help you grow your team.

Fill in your details and we’ll reach out to you within 24h.