Predicition.io and deep learning libraries

[Omer Ben-Ami <om...@roboro.io>]

Hi,
I'm new to prediction.io and trying to figure out what is the best way to
use it.
I'm mainly interested in building a scalable recommendation system that
will use some well known algorithms and a proprietary algorithm as well.

I could not find any information on what is the best practice for combining
predicition.io with deep learning (DL) algorithms.

For example, does it make sense to use DSSTNE? from the architecture
described here it seems the engine should simply wrap the communication
with the ECS: https://aws.
amazon.com/blogs/big-data/generating-recommendations-at-amazon-scale-with-
apache-spark-and-amazon-dsstne/

Of course the question is not specific to DSSTNE and relevant to any DL
algorithm the will perform better on GPUs.

Thanks,
Omer