Cloudera Data Science Workbench Reviews

We have different use cases. Our banking use case uses machine learning to identify customer life events and recommend the best-suited card products. These machine-learning models are deployed in our environment, where they run on a scheduled basis. We rely on the platform for every data science modeling. We have 170 models in production. It is very helpful for production and evaluation. 

I appreciate CDSW's ability to logically segregate environments, such as data, DR, and production, ensuring they don't interfere with each other. The deployment of machine learning is fast and easy to manage. Its API calls are also fast. 

The tool's MLOps is not good. It's pricing also needs to improve. 

The product is stable. 

The tool is scalable and works on a cluster and node system. 

If you don't configure CDSW well, then it might be not useful for you. Deploying the tool can vary in complexity, but most of the time, it's relatively simple and straightforward. Triggering a job from data to production is easy, as the platform automates the deployment process. However, ensuring optimal resource allocation is essential for smooth operations.

The product is expensive. 

We can integrate the product using APIs to connect with other systems. Additionally, integration with GitLab is straightforward, as it offers features to connect with external systems seamlessly. The product has enterprise customers and focuses on stability and performance. I rate it six out of ten. 

Snowflake has a lot of good things compared to the product. CDSW is on-prem, while Snowflake is on the cloud. Snowflake is better in terms of data lineage, data catalog, and pricing. 

I am a professor and this is one of the solutions that I use as a teaching tool for my students.

The most recent version can be used by the students while they are working in the labs because our hardware has sufficient RAM to run it. However, many of the students have laptops that do not have enough RAM, so they cannot work with it outside of the labs. In this case, the students are able to use an older version with lesser demand on hardware requirements.

The Cloudera Data Science Workbench is customizable and easy to use.

Running this solution requires a minimum of 12GB to 16GB of RAM.

In the future, I would like to see a student version of the Data Science Workbench that includes sample datasets that can be used for teaching purposes.

I have been working with Cloudera for a few years.

It is a stable product.

We found this solution a little bit difficult to scale. When we wanted to run it as a multi-cluster we had some issues, which is what made it difficult. We have approximately 200 students who use Cloudera to develop programs and complete projects.

I am satisfied with the technical support from the vendor.

This solution is very easy to install. In fact, all you do is open it in the virtual machine.

I would rate this solution an eight out of ten.