Amazon Deep Learning AMIs vs. Azure Databricks

Suitable: 1. Best for quickly setting up an instance with pre-installed libraries. 2. Ideal for people in Deep Learning space who struggle with Cuda / Nvidia driver installations. Not suitable: 1. People who want to install custom libraries or different version of those. 2. In these cases, updating the version of libraries many times leads to version mismatch which can cause many errors.

Incentivized

Suppose you have multiple data sources and you want to bring the data into one place, transform it and make it into a data model. Azure Databricks is a perfectly suited solution for this. Leverage spark JDBC or any external cloud based tool (ADG, AWS Glue) to bring the data into a cloud storage. From there, Azure Databricks can handle everything. The data can be ingested by Azure Databricks into a 3 Layer architecture based on the delta lake tables. The first layer, raw layer, has the raw as is data from source. The enrich layer, acts as the cleaning and filtering layer to clean the data at an individual table level. The gold layer, is the final layer responsible for a data model. This acts as the serving layer for BI For BI needs, if you need simple dashboards, you can leverage Azure Databricks BI to create them with a simple click! For complex dashboards, just like any sql db, you can hook it with a simple JDBC string to any external BI tool.

Incentivized

• Setting up environment
• Support for different types of machines
• Perfect for Machine Learning / Deep Learning use cases
• Nvidia / Cuda / Conda support easily

Incentivized

• Unity Catalog
• Data Federation in Lakehouse Architecture
• Integration of Mosaic AI in the SQL Layer

Incentivized

• Some aspects of the User Interface are quite confusing and activating packages can be a bit convoluted
• It can be a bit confusing to switch between frameworks for novice users

Incentivized

• Their pipeline workflow orchestration is pretty primitive. Lacks some common features
• Workspace UI and navigation requires steep learning curve
• Personally, I am not fond of their autosave feature. Its dangerous for production level notebooks scripts

Incentivized

Great for what we use day to day and does what we need it to do. Cost management is not fully developed across the UX and gets expensive very quickly for developing projects. Integrated very well with our Microsoft stack and can be worked on collaboratively which works well for us.

Incentivized

Both of these services provide similar functionality and from my experience both are top class services which cover most of your needs. I think ultimately it comes down to what you need each service for. For example Amazon DL AMIs allows for clustering by default meaning I am able to run several clustering algorithms without a problem whereas IBM Watson Studio doesn't provide this functionality. They both provide a wide range of default packages such as Amazon providing caffe-2 and IBM providing sci-kitlearn. My main point is that both are very good services which have very similar functionality, you just need to think about the costs, suitability of features and integration with other services you are using.

Incentivized

I have found Azure Databricks to be much better than Snowflake for handling bigger, diverse data types. Snowflake is much simpler and better for smaller warehousing. The real time processing is much better in Azure Databricks and we have much more language options. Snowflake is more expensive but simpler to use. Both are great for different needs.

Incentivized

• It has made our Data Science/ Machine Learning Courses easier to manage/ need less human input therefore allowing us to increase the cohort size for this degree
• It has unified a lot of technologies reducing the load on our IT team

Incentivized

• Helped reduce time for collecting data
• Reduced cost in maintaining multiple data sources
• Access for multiple users and management of users/data in a single platform

Incentivized