EnglishDeutschFrançaisEspañolPortuguês

Snowflake · SF-DS · Advanced

SnowPro Advanced: Data Scientist (DSA-C03) — Practice Questions and Mock Exam

The SnowPro Advanced: Data Scientist Certification (DSA-C03) validates the ability to apply data science principles within Snowflake, including defining data science concepts, performing data and feature engineering, building and deploying machine learning models, leveraging Snowpark for ML workflows, and presenting results through data visualization.

Practice for freeTry a practice question first
Try a practice question firstPractice for free
Start practicing SF-DS freeBuy SF-DS access for $29.99
65Questions per Pass-IT mock exam
115minPass-IT mock exam time limit

Official SF-DS exam: 750 out of 1000 points to pass.

Checked against Snowflake · September 2026 · Current exam version

Try Pass-IT

See what learning with Pass-IT is like.

Answer a SF-DS sample question and read the explanation with its source. Right here, with no sign-up.

Model Deployment1 / 5

When deploying a model via the Snowflake Model Registry, which method runs inference on a DataFrame?

AlexFull explanation from Alex

The ModelVersion.run() method is the correct API for running inference on a DataFrame through the Snowflake Model Registry. After retrieving a model version via Registry.get_model().version(), calling run(df) executes the model's prediction logic and returns results as a DataFrame. predict() resembles scikit-learn's API but is not a ModelVersion method. score() and execute() are also not valid ModelVersion methods. The run() method abstracts away model deserialization and provides a unified inference interface regardless of the underlying ML framework (scikit-learn, XGBoost, PyTorch, etc.).

Sourcedocs.snowflake.com

Data Science Concepts3 / 5

Which of the following best describes the F1 score?

AlexFull explanation from Alex

The F1 score is the harmonic mean of precision and recall: F1 = 2 * (P * R) / (P + R). The harmonic mean penalizes extreme imbalance—both precision and recall must be high for a high F1. The option “The arithmetic mean of precision and recall” (arithmetic mean) would be (P + R) / 2, which doesn't penalize imbalance as strongly and would overestimate performance. The option “The weighted average of all classification metrics” (weighted average of all classification metrics) is too broad and not the definition of F1. The option “The geometric mean of accuracy and recall” (geometric mean of accuracy and recall) uses the wrong metrics and wrong aggregation function. F1 is particularly useful for imbalanced classes where accuracy alone is misleading. Precision = TP/(TP+FP), Recall = TP/(TP+FN).

Sourcedocs.snowflake.com

Model Development5 / 5

A fraud team compares two decision thresholds for a scored validation set of 10,000 transactions. Each caught fraud saves 400 dollars, each false alarm costs 25 dollars in manual review, and missed fraud and correctly passed transactions cost nothing extra. Threshold TP FP FN TN A 80 120 20 9780 B 60 40 40 9860 Which threshold gives the higher expected payout, and what is it?

AlexFull explanation from Alex

A confusion matrix becomes a business decision once each cell has a value. Multiply the counts by their payoffs and add them up: here true positives earn 400 dollars, false positives cost 25, and the other cells are zero. Threshold A catches more fraud at the price of more reviews and nets 29,000 dollars, while threshold B nets 23,000. The tempting mistakes are to count the savings and forget the review costs, which gives 32,000 or 24,000, or to prefer the threshold with fewer false alarms without pricing them. scikit-learn's threshold tuning guidance makes the same point: a fixed 0.5 cut-off is rarely the best choice when error costs differ. Exam tip: write payout = value x count for every cell, then compare the totals.

Sourcelearn.snowflake.com

How to prepare for SF-DS with Pass-IT.

Currently 360 practice questions for SF-DS, with new ones added regularly. Focus on what you still need to learn.

Get past difficult questions.

Understand why an answer is correct. Alex explains the concepts and answers your follow-up questions so you can apply what you learn to the next task.

Meet Alex →

Try the available questions in the preview above.

Make your study time count.

Practice at your level.
For new practice sessions, Pass-IT adjusts the difficulty based on your previous answers.

Learn from mistakes.
Incorrect answers automatically become flashcards. When you review them, you rate how confidently you remember; Pass-IT schedules the next review.

See how focused practice works →

Try the available questions in the preview above.

Practice for the moment that counts.

Take a full mock exam with a time limit and find out whether you can finish in time.

Your SF-DS mock exam

Questions
65
Minutes
115

Illustration of an exam session

  1. Start
  2. Answer questions
  3. Finish

Length and time limit of the Pass-IT mock exam.

Go into your exam with more confidence.

Your Readiness Score helps you assess whether you are ready for the exam. Results by exam domain show you where you should keep practicing.

Your money back if you fail despite 80% Readiness.
If your Readiness Score is at least 80% when you take the exam and you still fail, we refund the Pass-IT purchase price for this certification.

For active, paid access. Submit proof of failing within 14 days.

Further guarantee conditions

You need at least 80% Readiness and active, paid access to the certification when you take the exam. Take the exam within your access period and submit official proof of failing within 14 days after the exam.

For purchases through Apple or Google, refunds are handled by the respective store. See the full terms for further details.

Read the guarantee terms →

The Readiness Score describes your practice progress, not a probability of passing.

Guarantee requirement

80%

At least 80% Readiness
when you take the exam

Money back if you fail

Question counts and mock exam settings from the current Pass-IT catalog.

Exam domains

What you practice for SF-DS.

Exam topics, relevant practice and sources you can check.

Frame the data science problemSee the guide for weighting61 QuestionsPass-IT questions

Choose learning methods and statistical tools that fit the question.

  • Distinguish supervised, unsupervised and reinforcement learning, and match regression, classification, forecasting, segmentation, clustering or generative AI to the problem.
  • Explain the lifecycle from collection and exploration through feature engineering, training, deployment, evaluation, monitoring and model versioning.
  • Interpret normal or skewed distributions, outliers, the central limit theorem, Z and T tests, bootstrapping and confidence intervals.
Official Snowflake objectives
Prepare data and useful featuresSee the guide for weighting99 QuestionsPass-IT questions

Turn raw data into inputs that a model can learn from.

  • Clean and sample data with SQL and Snowpark for Python, using joins and aggregates while handling duplicates, irrelevant fields, missing values and types.
  • Explore patterns with SQL, Snowpark and external notebooks; calculate descriptive statistics and interpret regression slopes, intercepts and variable relationships.
  • Scale, normalize, encode and bin features, create derived attributes with DataFrames and manage features in Snowpark Feature Store.
  • Use Snowsight, Notebooks and graph libraries to present statistical summaries, identify outliers and explain the business case.
Official Snowflake objectives
Develop and evaluate modelsSee the guide for weighting106 QuestionsPass-IT questions

Build models and examine how well they solve the intended problem.

  • Connect Python, Snowpark ML, pandas and external IDEs to Snowflake data, and recognize the supported Snowpark languages.
  • Use Cortex embeddings, prompting, fine-tuning and task-specific models for generative AI and language tasks.
  • Build automated training pipelines with dynamic tables, Python UDFs, UDTFs and procedures or external functions; tune hyperparameters and choose metrics, cross-validation, holdouts and resampling.
  • Validate classification or regression with ROC curves, confusion matrices, residual plots and metrics, relating results to expected business value.
  • Explain predictions through feature impact, partial dependence, confidence intervals and SHAP values.
Official Snowflake objectives
Deploy and maintain modelsSee the guide for weighting94 QuestionsPass-IT questions

Put models into production and detect when their performance changes.

  • Serve external models through external functions or deploy in Snowflake with Python UDFs, stages, Model Registry and Snowpark Container Services, storing predictions as needed.
  • Compare production and training data to detect drift or model decay, and evaluate AUC, accuracy, precision, recall or RMSE to decide when retraining is needed.
  • Tag model metadata, manage versions in Model Registry and automate retraining as part of the model lifecycle.
Official Snowflake objectives
Question formats you can practice
Multiple Choice
Select one answer.
Multiple Response
Select multiple applicable answers.

These are Pass-IT practice formats. Check the Snowflake certification page for the official exam information.

Frequently asked questions

Your preparation

How much study time should I plan for SF-DS?

For planning, Pass-IT estimates roughly 50–120 hours. Relevant experience may put you toward the lower end; starting from scratch may put you toward the higher end. Your needs may fall outside this range. This is not a time requirement from Snowflake or a promise that you will be exam-ready within that time.

What do you need before the Advanced Data Scientist exam?

SnowPro Core certification, plus two or more years of hands-on Snowflake data science work in production according to Snowflake's own guidance. The catalog budget is around 80 hours. The Core requirement is enforced, so it has to be passed first regardless of how much modelling experience you bring.

Using Pass-IT

What language are Pass-IT questions in?

Questions in the app are currently in English. Translations of the sample questions on this page apply only to this preview. Check with Snowflake for the languages available for the official exam.

How long does my SF-DS access last?

Access covers this certification for 12 months. You pay once, with no automatic renewal.

When does the money-back guarantee apply?

The guarantee refunds the Pass-IT purchase price for this certification. You need eligible paid access and a Readiness Score of at least 80% from your practice results when you take the exam. This score can only reach 80% after at least 50 answered questions and is not a validated probability of passing. Take the exam during your access period and submit official proof of failing within 14 days. Guarantee terms

About the SF-DS exam

Where can I take the SF-DS exam?

Online proctored or onsite testing centers.

How long is the SnowPro Advanced: Data Scientist (DSA-C03) exam?

The SnowPro Advanced: Data Scientist (DSA-C03) exam has 65 questions and a 115-minute time limit.

Advanced Data Scientist or Advanced Data Engineer?

The data scientist exam puts 31% on model development and 25% on deployment, so it is about building and shipping models inside Snowflake. The data engineer exam puts 28% on data movement and 25% on transformation, so it is about the pipelines that feed them. Both require SnowPro Core and neither ranks above the other.

How long does the Advanced Data Scientist certification last?

Your certification remains valid for two years from the date you earned it, including after the exam’s retirement effective 1 November 2026. To renew through continuing education, complete one eligible Snowflake instructor-led training course or earn an equivalent or higher-level SnowPro certification before yours expires. For renewal, Snowflake recommends considering the SnowPro Advanced: MLOps Engineer Certification.

Your preparation for SF-DS

$29.99one-time

12 months of access for SF-DS · No automatic renewal

Buy SF-DS access

Questions in the app are currently in English.

Try 20 questions for freeFree account required. No credit card.

Already have an account? Log in

Prefer a subscription? View subscriptions →

Money back if you fail with at least 80% Readiness, for eligible purchases under the guarantee terms.

Guarantee terms

The guarantee refunds the Pass-IT purchase price for this certification. You need eligible paid access and a Readiness Score of at least 80% from your practice results when you take the exam. This score can only reach 80% after at least 50 answered questions and is not a validated probability of passing. Take the exam during your access period and submit official proof of failing within 14 days.

Full guarantee terms →

Related certifications

SnowPro Advanced: Administrator (ADA-C02)SnowPro Advanced: Architect (ARA-C01)SnowPro Advanced: Data Analyst (DAA-C01)SnowPro Advanced: Data Engineer (DEA-C02)All Snowflake certifications →