My Reflection on GCP Professional Machine Learning Engineer Certification Exam

If you had taken the GCP Professional Data Engineer Certificate, you will find the two have many overlaps in concepts. The 4-hour, 120-question exam put majority of emphasis (more than ~70%) on non-modeling topics. That means it would not be a ML knowledge test on how to make the best model with Tensorflow, but rather more on how to “engineer” the best ML solution given differnt constraints (time, money, etc). In that sense, it looks similar to the general Professional Data Engineer exam in the line of thinking, but focuses exclusively on ML solutions at its core.

Given the dearth of information at the time of preparation, I spent most of time with the coursera courses and docs. The first course offers a high level overview of running ML in GCP. The second specialisation introduces the nuts and bolts of doing ML with TensorFlow in GCP. The third specialisation delves into productionising ML and discusses in depth on improving ML models. All of them are relevant to the exam questions except for very specific TensorFlow functions and techniques that are presented in the last course for understandable reasons.

In addition to the online learnings, I also found topics below quite important — both in preparing for the exam and possibly the day-to-day job as a ML engineer.

3. For a given run, the Pipelines Lineage Explorer(Beta) shows the history and versions of your models and data.

To summarise, a viable plan to study for the exam I’d recommend would be Coursera -> docs -> tutorial. If you are running out of time, then focus on key ideas for the first two coursera MOOCs and the first two courses from the third coursera specialisation (End-to-End Machine Learning with TensorFlow on GCP and Production Machine Learning Systems), then read through documentations on relevant GCP products on machine learning, which some of them can be found below.

Related posts:

If you want to reduce financial stress, you might be focusing on reducing your outgoings, increasing your income, and paying down debt. All admirable aims, but in the long-term, some of the things…

At the end of the 20th century, observations done by NASA of supernovae millions to even billions of light years away using the Hubble Space Telescope showed that the universe’s rate of expansion in…