NVIDIA-Certified Professional: Accelerated Data Science

The exam is professional-level and requires 2-3 years of hands-on experience with GPU-accelerated data science. It covers 6 domains including data manipulation with RAPIDS, MLOps, GPU computing, and machine learning. Candidates should have strong practical experience with cuDF, cuML, cuGraph, and Dask.

NVIDIA does not publicly disclose the exact passing score. The exam contains 60-70 questions and candidates have 120 minutes to complete it. We recommend scoring 70%+ consistently on practice tests before scheduling.

The NCP-ADS certification is valid for 2 years from the date of issuance. To maintain your certification, you must retake and pass the exam before it expires.

NVIDIA recommends several courses: 'Accelerating End-to-End Data Science Workflows' ($90, 8 hours), 'Accelerating Clustering Algorithms' ($30, 2 hours), and instructor-led workshops including 'Fundamentals of Accelerated Data Science' and 'Enhancing Data Science Outcomes' ($500 each, 8 hours).

No prerequisite certifications are required, but you should have 2-3 years of experience with GPU-accelerated data science tools and workflows.

You should have hands-on experience with NVIDIA RAPIDS (cuDF, cuML, cuGraph), Dask for distributed GPU processing, XGBoost GPU training, Docker/Conda environments, DLProf profiling, and Python data science libraries.

Yes, this is an official NVIDIA Professional certification validating advanced GPU-accelerated data science skills. Upon passing, you receive a digital badge and optional certificate. GPU data science expertise is increasingly valued as organizations adopt accelerated computing.

Yes, the exam is delivered online and is remotely proctored via the Certiverse platform, allowing you to take it from anywhere with a stable internet connection.

The exam costs $200 USD. You'll need to create a Certiverse account to register for the exam.