The beginning of August was relatively productive for me – I decided that PL-300 and GCP ADP certification would be not enough for this summer, and decided to receive a bit more formal badges as a proof, that I know something. I’m still not sure that the fact of receiving of certificate after proctored exam means anything – but anyway, no one can stop me in this journey!
This time, I focused on two paths – Databricks Data Analyst and GCP Gen AI Leader. Both certifications are not so challenging, but at the same time cover relatively interesting topics, so let me briefly tell about me experience.
Databricks Data Analyst
I heard a lot about Databricks from data engineering community (especially in terms of Azure stack), but, unfortunately, my personal experience with it was pretty limited – ok, I knew that it is data platform which can be placed inside of cloud; that it makes life o data engineer easier; and that it is more straightforward comparing to impassable jungles of ETL tools inside of GCP, Azure or AWS family. But if you ask me about the difference between Databricks and, for example, Snowflake, I leave just silence after me.
That is why I decided to fill some gaps in that, and proceeded to Databricks Academy. Data Analyst route (https://www.databricks.com/learn/certification/data-analyst-associate) contains 4 major steps for completion, which on the first glance looks relatively simple:
- Databricks fundamentals course – about lakehouse as a concept, medallion architecture, delta lakes, integrations, the most major products inside of platform;
- BI/AI Analytics capabilities course – interesting that Databricks unites BI and AI in one section, and in this course authors are focusing on different use cases on how AI (either in notebook or in Genie)can help with reporting activities;
- SQL on Databricks course – it uses ANSI, so it is not significantly different from what I saw previously, and also provides superficial overview of usage of Spark and UDFs;
- Databricks Data Analyst proctored exam as is.
It was really interesting! Especially I liked a part about SQL, because in that part authors described in details the structure of Databricks projects and how it works.
The overall preparation time for exam took for me ~1 week, or even a bit less. Outside of videos on Databricks, I had a chance to touch free version of platform, and to complete several free mock tests. So, I didn’t have any additional trainings or sessions with mentor – to be honest, the exam in current format is too simple to spend time and money on such activities.
Interesting thing about exam – ~50% of questions are related to SQL. Of course, there are not questions about the difference between WHERE and HAVING, but much more about window functions, UDFs, CTEs, subqueries – but it is still just SQL. It means, that if you already had a chance to work on any other SQL-driven solutions at least for couple of months, preparation for Data Analyst exam will be not hard at all – you can just check the main concepts of the platform (based on courses material), check UI, and you already will be partially ready.
So, I passed it with the first attempt. I cannot say it was completely challenging – most difficulties I experienced with Data Management-related questions, but it was compensated by other sections. Also interesting that it was not so much questions about AI part of Databricks – but maybe it will be changed in future.
Google Cloud Platform Gen AI Leader
Gen AI leader is the most fresh certification on GCP – and the second foundation-level one. It means that it is not so much aimed on tech-related part of LLMs, but significantly more on end-user experience.
Gen AI Leader learning path (https://cloud.google.com/learn/certification/generative-ai-leader) mostly focuses on basic concepts of LLM consumption (prompt engineering, RAG, grounding, etc.), and introduces a variety of AI/Ml-related products hosted on GCP (Vertex AI, NotebookLM, Gemini, Veo, Imagen). Learning path can be completed in just one day – course modules contain just 5-minutes short videos and some visual materials, most of them are relatively light for non-tech audience. Especially I liked that each video ended with phrase “Ready to learn? Let’s dive in!” – by the end of learning path it started working as a mantra 😀
I cannot say that I actively use LLMs in my daily life, especially Google products, but preparation for certification exam took for me just 2 days. This time I even didn’t try to complete any mock exams (only checked 5-10 practice questions from Google guide), and proceeded to exam completion almost immediately after learning courses completion.
Questions on exam were a bit similar to GCP ADP certification – in almost each case authors provide a use case, and you need to select right solution from the perspective of appropriate Gen AI techniques or Google products. I would say that practice questions are pretty representative comparing to real exam.
So, as a result, I have two more papers no one takes care of! I’m not sure that it makes any sense, but at least it was fun!
Leave a Reply