Databricks Data Intelligence Platform
Databricks, Inc.External reviews
637 reviews
from
and
External reviews are not included in the AWS star rating for the product.
Rapidly Improving AI Platform With Flexible Tooling
What do you like best about the product?
Scheduled jobs, pre-installed ML environments, abundance of documentation and example use cases. There are a lot of best practices available through the Spark AI conference also.
What do you dislike about the product?
Their Spark install is not always compatible with 3rd party tools (ex: geospatial) causing a delay in some package availability until an extra release. Some gotachs with distributed model training (Horovod).
What problems is the product solving and how is that benefiting you?
We are training models to forecast weather and have been able to achieve scale without having to do create a custom architecture. We now have experimentation happening 10X faster than before.
Databricks
What do you like best about the product?
I really like that we can have queries. We have only started to review them properly for each account, but there are limitless opportunities once we start digging.
What do you dislike about the product?
I would like to access more access to creating queries and investigating several things. Sometimes I get glitches with the account, it freezes and I can't see the information.
What problems is the product solving and how is that benefiting you?
We are finding fraudsters frequently and once we track a trend, we can solve them much easier. The best part is finding out information about a customer's account that we cannot see on the other system we use
Great customization and granularity of data
What do you like best about the product?
I like how granular we can get when running queries and reports for our business purposes
What do you dislike about the product?
The user interface could be a little sleeker and more intuitive
What problems is the product solving and how is that benefiting you?
Monitoring various categories of data input
Databricks allows us to pull Canadian Credit file information to verify our Customers.
What do you like best about the product?
I like how Data bricks is layed out. At first it can seem difficult to read, but from computer background i find it a great way to layout the information in an organized fashion.
What do you dislike about the product?
Sometimes Credit Files will pull not pull in the Credit Kyc section and will almost always pull in the Customer information Tab.
What problems is the product solving and how is that benefiting you?
We are abel to identify Customers as we onboard them onto our business and ensure they are legitimate individuals and mitigating risk.
Amazing Platform for an Enterprise Data Warehouse and Data Science Platform
What do you like best about the product?
Incredible support for Data Science, Machine Learning and SQL use cases all in one platform.
What do you dislike about the product?
Theoretically their SQL Engine may not scale, but I have not seem that at all.
What problems is the product solving and how is that benefiting you?
Predictive outcomes on cases, lead generation, optimization of marketing strategies and support for common tools like PowerBI.
Recommendations to others considering the product:
Why use proprietary APIs when you have such a complete and open solution with DataBricks.
Good but UI is not perfect
What do you like best about the product?
The ability to monitor and relaunch jobs from the phone is amazing. I don't need to be on my computer on the weekend to relaunch a failed job.
What do you dislike about the product?
I don't particularly appreciate how the mlflow menu is hidden from the left bar UI unless I enter a specific link that contains the ID of an experiment.
For instance, I cannot see anything in https://*******.cloud.databricks.com/#mlflow and I cannot see mlflow in the left bar unless I go to :
https://*******.cloud.databricks.com/#mlflow/experiments/3817218/runs/172f6fcb9f144bbc93cc5b3b857c50f1
For instance, I cannot see anything in https://*******.cloud.databricks.com/#mlflow and I cannot see mlflow in the left bar unless I go to :
https://*******.cloud.databricks.com/#mlflow/experiments/3817218/runs/172f6fcb9f144bbc93cc5b3b857c50f1
What problems is the product solving and how is that benefiting you?
We are using databricks for weekly ETL and neural network model training and tracking of results. We like the fact that we can use different cloud providers based on our needs.
Easy of use
What do you like best about the product?
The easy of use is the most useful feature i like about MLFlow, it can use locally without any need for deployment on any server, great UI which allows use to search through general experiments and flexibility of changing data stores.
What do you dislike about the product?
The UI design is in Django I think which is a bit laggy and slow improvement can be done on that side.
What problems is the product solving and how is that benefiting you?
We tried to provide our data store as a plug-in for MLFlow.
Recommendations to others considering the product:
Yes
MLFLow as a great addition to our ML focused data pipelines
What do you like best about the product?
MLFlow has been instrumental in providing a decoupled interface between training and prediction part of our ML pipelines where we can store model metadata as part of training cycle and then in prediction we are able to leverage and pick model with highest ROC or just latest by chronological sorting and overall just do good job in tracking our experiments.
What do you dislike about the product?
Sometimes slow to render in databricks environment as UI but that could be related to our databricks setup.
What problems is the product solving and how is that benefiting you?
MLFlow is very helpful in ability to do concurrent run of training and prediction as prediction doesn't have to wait for training to finish, we can just pick last successful experiment from MLFlow and leverage it for predictions.
Great tool, helped me alot.
What do you like best about the product?
Centralization and remote servers and API
What do you dislike about the product?
Nothing specific. I like everything about mlflow.
What problems is the product solving and how is that benefiting you?
experiment tracking and ML metadata
Great for model refresh
What do you like best about the product?
Great for model refreshes and comparison
What do you dislike about the product?
Ui and artifact loads can be improved...
What problems is the product solving and how is that benefiting you?
Model storage, reuse.
Recommendations to others considering the product:
Great for storing your artifact.
showing 261 - 270