Skip to main content

Professional Cloud Data Engineer

🌸 Passed: December 28, 2024

Related:

Exam - December 28, 2024 - Passing Memo​

Key Points Memo After Passing

Reference Links

Exam Points

  • The business case study questions were removed.
  • Many advanced questions on windowing.
  • Many system-level questions about Dataproc.
  • Networking, deployment automation, and security topics also appeared frequently in conjunction with data services.

Notes on Dataproc

  • I want to clarify whether Dataproc's intermediate system utilizes on-premises or memory.
Exam Information - December 28, 2024

Reason for taking the exam: A coupon was expiring on that date, and this was the exam for which I had the most knowledge and the best chance of passing. Also, due to date constraints, I chose the online exam.

  • Exam Name: Google Cloud Certified - Professional Data Engineer (Japanese)
  • Date: December 28, 2024
  • Time: 09:15 AM
  • Exam Method: Remotely proctored online exam

Preparation:

  • Preparing to take it in the kitchen. It seems best to stick a sheet to the wall.
  • Download the secure browser (I'll do this at home).
  • Driver's license / Insurance card / Credit card.

🔥Strategy for the Exam🔥​

Prompt Memo for Creating Practice Questions

2024/11/30

  • I have configured a dedicated prompt in GPTs. However, when working, it is necessary to reinforce the prompt as needed. For example, reinforcing the latter part of a long generation or working on a new page for multiple prompts is necessary to compensate for the weaknesses of generative AI.

Final working prompt

  • In the worst-case scenario, I managed by including the question number and the question text in the instruction.
Regarding the questions in "Knowledge" "Chapter 4: Building and Operating Data Pipelines (Questions 151-200)",
please improve the next 5 questions starting from question 165 as instructed. Be sure to "search" the "Knowledge" for confirmation and correctly understand the correspondence between the before and after versions of the questions and the intent of the instructions (format and explanation improvements).
- The explanations should be detailed and helpful for a Google Cloud beginner.
- For important Google Cloud service names appearing in the explanation, write them as "Official Name (e.g., Cloud Spanner)".

``Reference info for searching the relevant question in "Knowledge"
Question 165
What is the most common concept used for state management in streaming processing with Dataflow?
``
GPTs Tips​

Most effective measure for long text generation:

  • If the output becomes garbled, create a new page and prompt again.
    • → This is more effective than improving the prompt, likely because it reduces the system load.

Second-best measure for long text generation:

  • Split the material to be registered in the knowledge base.
    • → This is presumably effective for modification and formatting tasks that follow the order and content of the original material.

Magic words (prompt) for when contradictions occur:

  • Be sure to "search" the "Knowledge" for confirmation.
    • → By making the GPTs search the registered material, it's possible to enforce generation that adheres to the format, content, order, and instructions of the reference material.

Standard best practice:

  • Be specific and precise with prompts.
    • → In GPTs, you can register instructions in advance, so aim for clear instructions at that stage.
    • → It's also important to supplement the pre-registered instructions on the work screen.
TIPS for Generating Advanced Questions​

2024/12/14

Improving Mock Exams: Copy the mock exam from the official website and save it as a text file. ↓ Register it in GPTs and organize the format. ↓ Gradually improve it in Gemini (Studio) after setting up a System Prompt.

Improving Basic Questions: Generate a basic question set based on the exam scope using GPTs. ↓ Improve them into advanced questions in Gemini (Studio) after setting up a System Prompt.