H2O.ai Reviews and Pricing

it_user862530

Associate Consultant at a tech services company with 201-500 employees

May 2, 2018

Download

AutoML helps in hands-free evaluations of ML algorithms, but solution needs a GUI

Pros and Cons

"AutoML helps in hands-free initial evaluations of efficiency/accuracy of ML algorithms."

"It needs a drag and drop GUI like KNIME, for easy access to and visibility of workflows."

What is our primary use case?

Testing/modeling data in the initial stages of approaching a machine-learning problem. Environment: Laptops running Ubuntu 16.04/Python 3.

What is most valuable?

AutoML helps in hands-free initial evaluations of efficiency/accuracy of ML algorithms; with training input data.

What needs improvement?

It needs a drag and drop GUI like KNIME, for easy access to and visibility of workflows.

For how long have I used the solution?

Less than one year.

Disclosure: My company does not have a business relationship with this vendor other than being a customer.

it_user837546

Principal Data Scientist

Apr 5, 2018

Download

Provides fast training, memory-efficient DataFrame manipulation, well-documented and easy-to-use algorithms

Pros and Cons

"Fast training, memory-efficient DataFrame manipulation, well-documented, easy-to-use algorithms, ability to integrate with enterprise Java apps (through POJO/MOJO) are the main reasons why we switched from Spark to H2O."

"Referring to bullet-3 as well, H2O DataFrame manipulation capabilities are too primitive."
"It lacks the data manipulation capabilities of R and Pandas DataFrames. We would kill for dplyr offloading H2O."

What is our primary use case?

We currently use H2O for real-time predictive analytics for fraud prevention. We have a Java-based feature engineering pipeline that attaches to POJO objects obtained from H2O. We run the whole pipeline on lightweight VMs and get under 1ms latency for each real-time transaction scoring.

How has it helped my organization?

We previously needed a four-machine Spark cluster to be able to train an ML model using tens of millions of transactions, and hours of time during the modeling phase. Currently, same training can now be done on an old MacBook pro with 8GB RAM within few minutes.

What is most valuable?

Fast training
Memory-efficient DataFrame manipulation
Well-documented, easy-to-use algorithms
Ability to integrate with enterprise Java apps (through POJO/MOJO)

These are the main reasons why we switched from Spark to H2O.

What needs improvement?

Referring to bullet-3 as well, H2O DataFrame manipulation capabilities are too primitive.

For how long have I used the solution?

One to three years.

What do I think about the stability of the solution?

Yes, we ran into a few bugs and opened JIRA tickets with reproducible test cases. We found workarounds for the problems ourselves.

What do I think about the scalability of the solution?

No issues with scalability, it works smoothly.

How are customer service and technical support?

We use the open-source/community branch and get support through forum discussions.

Which solution did I use previously and why did I switch?

We used to developing on Scala + Spark ML. We switched, at least in part, due to reasons mentioned in the Valuable Features section of this review.

How was the initial setup?

Initial setup is very easy through pip, JAR download, or R install.packages.

What's my experience with pricing, setup cost, and licensing?

Currently, we do not purchase enterprise support.

Which other solutions did I evaluate?

We have experience with pretty much everything available; hence, the switch was an informed decision and natural.

What other advice do I have?

We rate it at eight out of 10. It is very fast, light-weight, well-documented, and low-maintenance. The reasons it is not rated 10 are, it lacks the data manipulation capabilities of R and Pandas DataFrames. We would kill for dplyr offloading H2O.

Disclosure: My company does not have a business relationship with this vendor other than being a customer.