A new piste to ski: Taking the SnowPro Advanced Data Scientist Certification Exam

Me skiing in Davos, Switzerland a couple of years ago

I don’t consider myself as an advanced skier as living in Texas, I get a chance to ski once or twice a year only. However, I am passionate about skiing, exploring new mountains/ski slopes (“pistes”), even some black diamond ones, knowing that I might not be able to complete them. It is all about getting out of my comfort zone!

As a Solutions Architect and a Data Engineer, I do not consider myself as a Data Scientist as I am not a Data Scientist by training (if we don’t count the Coursera Data Science courses I attended over the years); however, in the last 10 years, I’ve had the privilege to work on very interesting Data Science projects as an Architect with awesome Data Scientists that I learned a lot from. I am also well aware that there is a lot of potential for innovation and growth in the Data Science landscape in general and the future of Data Engineering entails convergence of Data Science and Data Engineering roles into one. With these thoughts, I decided to take the brand new SnowPro Advanced Data Scientist Certification beta exam with a relatively lower expectation of myself. I also felt like there would be some knowledge overlap between this certification exam and all the other SnowPro Advanced certification exams I successfully completed in the past.

First things first! All things about the SnowPro Advanced Data Scientist Exam can be found here. The target audience for this exam is expected to have “2+ years of practical data science experience with Snowflake, in an enterprise environment.” The exam has a combination of 70 multi-select, multiple-choice questions and you have 115 minutes to complete them. (I finished the exam ~15 minutes earlier so I had 15 minutes to go over the questions that I marked for review.) The Data Scientist exam is designed to test the following domains:

Having studied for other certification exams that contain AI/ML test questions in the past (e.g GCP Professional Data Engineer certification) plus my deep expertise in Data Engineering with Snowflake, I felt like I should easily get the ~60% of the questions correctly. I also spent a week studying for some of the model development and deployment topics listed in the Exam Topics section as well as brushing up my statistics knowledge by taking a Udemy Data Science course. Snowflake also provides some resources to study including an Instructor-led Course and a free self-study guide:

Data Science with Snowflake https://www.snowflake.com/

Overall, with the exception of a few techniques/terminologies that I never heard of, I was comfortable with 80–90% of the questions with my previous experience and one week of studying the concepts. (I am still waiting for my exam result at the time of writing this post. UPDATE: I passed the exam.)

In my test, there were a good number of questions around Data Science concepts such as model accuracy, confusion matrix, bias/variance tradeoff, overfitting/underfitting, and hyperparameter optimization and imputation techniques. Also please make sure that you are familiar with the common algorithms and when to use them. (https://towardsdatascience.com/10-machine-learning-algorithms-you-need-to-know-77fb0055fe) There are some use case questions (which are my favorite!) around Snowflake SQL aggregate functions as well as Snowflake Python/R/Spark connectors and the best practices around when and how to use these. There were also a good number of Data/Feature Engineering questions around building continuous data pipelines in Snowflake using Snowpipe, streams and tasks, zero-copy clone feature, and leveraging data sets in Snowflake Data Marketplace.

Data Science in the Snowflake Cloud is still evolving and there are a lot of exciting advancements and innovations coming up in the upcoming months. The certification exam also has questions around the latest and greatest features like Snowpark and Java UDFs including code snippets for testing your knowledge on model scoring, inference, and model deployment so it is really helpful to get hands-on experience with these newer features.

Regardless of my exam result, I also feel like taking the SnowPro Advanced data Scientist exam was an awesome learning experience overall. Overall, I thought the SnowPro Advanced Data Scientist exam covers a great combination of data science and data engineering topics to test your understanding of the comprehensive Snowflake Data Cloud vision: A unified and governed platform that supports all personas for modern analytics. And as always, with the right level of balance between ease of use and control.

I am currently a Snowflake Pre-Sales Architect at Snowflake. Opinions expressed in this post are solely my own and do not represent the views or opinions of my employer or Snowflake training department.

--

--

Eda Johnson
Snowflake Builders Blog: Data Engineers, App Developers, AI/ML, & Data Science

AWS Machine Learning Specialty | Azure | Databricks | GCP | Snowflake Advanced Architect | Terraform certified Principal Data Cloud Architect