On other courses there was a thread for exam experiences – I thought I’d share my experience of the GCP Professional Data Engineer.
My background: Business Tech consultant in AI/ML, Math background, GCP ACE cert
General comments: This course was a great prep guide. There was a a lot on BQ and BT, with a number of questions that asked about the nuts and bolts of performance and IAM policies. The business scenario questions were generally quite clear and as long as you understood the differences between the services they were manageable.
Strange Stuff: The questions of ML specifics seemed out of place with otherwise operational considerations. It’s worth making sure that you’re comfortable with the ML services and some of the detailed points on common techniques.
Kryterion were about 20 minutes late starting the exam, a little frustrating given you’re already in the software and I wasn’t sure if exiting out of it was allowed / would screw up the exam. Ended up contacting them on my phone after 15 mins. After the exam, I got the prelim result, but no notification other than the usual 7-10 business days. ACE turnaround was about 24 hours for me, so after 5 days I got a little nervous. Turns out that the switch to online proctoring has created a bit of a bottleneck in the verification process and they’re aiming for a 3 week verification turnaround.
Hi Guy, thanks so much for sharing this. And congratulations on passing your cert!
I passed mine yesterday (provisional result, waiting for the cert!). I had a nightmare as the sound card packed up on my laptop just as I was trying to log in, so had to hurriedly set up sentinel on another PC. I was about 10 mins late logging but Kryterion were 40 mins late starting the exam so it didn’t matter!
The questions were a real mix, maybe 10 I answered with no hesitation, 30 needing some thought and 10 I came back to several times. I did a first pass in only 40 minutes or so then stepped through them twice more just to review. Nothing particularly unexpected except for a couple of ML questions that threw me!
I found the A Cloud Guru course very useful for preparation as it gave some structure to what I needed to study and there is a lot of information packed in. There’s a few parts where the course could do with updating (e.g. SQL server out of beta for Cloud SQL, RAM and CPU limits increased here and there) but it was a solid grounding. I recommend going through the google docs as well, especially on the ML topics. I felt the cloud guru content could do with tightening up here maybe with some recorded demos to bring it to life. I probably studied about 30-40 hours in total though I had done some prior overview courses on pluralsight and coursera.
I also had technical issues when I first attempted passing the exam (remote proctored) a few weeks ago. My fairly recent Mac (model A2141) running the latest Catalina OS failed on me at the worst possible time! I had installed the Kryterion software the day before and successfully ran all the tests. At the time of the exam, I successfully established communication with the proctor, ready to take the exam. But before I had even started the exam, the Mac froze on me, and after about 30 seconds it abruptly rebooted on me. At the user prompt, I would log in again to my Mac, but it would keep rebooting on me during the boot process. I didn’t have another suitable machine to take the exam, so I couldn’t take it at that time.
I spend two days trying to revive the Mac, I even wiped out the partitions and reinstalled the OS from scratch, as per advice from Apple support. None of that could revive the Mac, and we therefore believe that the device had a hardware failure. I’m finding very strange coincidence that I had such failure at this crucial time when I was just about to take the test. Anyone else had a similar issue?
According to Kryterion technical support, this Mac device has a "virtual OS" (so they said), and hence it is not supported. I hadn’t seen anything about that in the system requirements. Kryterion pointed me to Google to see if they could grant me a voucher to take the exam another time. It took some back and forth discussing with Google what had happened, but in the end they did grant me a voucher.
So I kept studying for some extra weeks so that I’d stay sharp for the exam. This time, I didn’t want to have more technical problems. So I registered for an onsite exam.
Last night, I reviewed all the exam tips from the course and as well dug deeper into some of the sections that I were less at ease with. I did once more both practice exams from the course and from Google. A few of those questions were similar on the exam today, and I’m happy to report that I passed!
Thanks very much for the great course content! After a little breather, I will likely continue with more courses.
I cleared GCP data engineer certification on 17th Oct. Waiting for official mail from Google. Just wanted to share few tips which helped me to crack the certification. Its quite tough and tricky exam. Understanding concepts of each services is a must to clear the exam
I took following courses
1. Coursera ( sponsored) – There are 5 courses which needs to be completed but these courses are not sufficient to clear the exam. They just give you good concepts
2. ROI training (sponsored) – brushed through them at high level but the practice tests were not of great standard
3. Purchased ACG data engineer course – Excellent and very well explained . Mock tests give you a sense of what type of questions appear but nothing similar came up on the exam.
4. GCP official guide by Dan Sullivan. – Fantastic guide. Read this book end to end and took practice tests at end of each chapter. This is a must read .
5. Google documentation : Windowing concepts, Big table and Big query ( read end to end as lot of questions came from BQ and big table)
Practice tests which I took
Udemy tests and ACG tests gave a good sense of quizzing on concepts which helped in the exam. I took these exams repeatedly till I got above 90% . Go through each explanations for all questions.
Questions split :
Big Query – approx 10 questions ( tricky questions)
Big table – around 5
Datastore – 3 to 4 questions
Tensorflow and ML – around 20 questions ( tricky questions here)
Spanner – around 3 questions
Storage concepts ( GCS) – 5 to 6 questions
Security Keys – around 2 questions
You don’t need deep handson lab practice just to pass the exam but its a must if you want to pursue career in GCP .
I took 1 hr for first pass and had marked around 20 questions for review. did 2nd pass in 30 min and gave a third pass in 15 min.. just 8 minutes were remaining before I submitted the exam.
All the best !!