Advancing Spark - Introduction to Featurestores

[@phillipbadenhorst8069]

The biggest advantage for most companies sits in having a feature store feed BI, advanced analytics and ML. Using it to speed up productivity and improve consistency in analytics is hugely beneficial.

[@jist3349]

Nice video, I personally see huge added value of feature store. A few notes - you can see 20+ companies they are working with that e.g. #iguazio #hopsworks #feast #tecton #michelangelo #databricks #kaskada #rasgo #sagemaker #vertexai #molecula #abacusai #scribbledata #h2o ... and second note when you join this one with MLOps and view to the future such as model monitoring, adaptable AI, ... you can extend your view.

[@fawzirida2233]

What is the difference between storing these engineered features in a Feature store over a Gold table (in a Lakehouse architecture)?

[@jamesang8735]

My question precisely. Could someone share his thoughts? My boss is recommending to store the data in feature store as gold layer before feeding it into dashboard in power bi. My initial idea was to store the data in a delta table as gold layer. What are the pros and cons of both methods?

[@houmandehesh740]

This was AMAZING thank you so much Simon, as always 🙏

[@bigchallenge7987]

I used the Hopsworks feature store and experienced its capability. That's quite amazing! But really excited to use data bricks feature store which is in preview mode now:)

[@charlesdelanoy]

Thanks for this, insightful, I appreciate all the videos. Please keep going.

[@gabrielfaria7841]

Once again great work and clear explanation!

[@jamesang8735]

I love both of your smile. It makes my day. =)

[@climbit9555]

for us the DB Feature Store is quite usefull. But still it would be nice to see someone else using it :)

[@mimmakutu]

Just for completeness, aws launched a featurestore as part of sagemaker suite

[@hubert_dudek]

Smaller squares on shirt and bigger behing Gavy. I think feature store is nice to store weather temperature but not necessary milion of whether satellite photos.

[@jist3349]

Feature Store is not replacement of datalake ... it has really different usage

[@aromn]

Perhaps the name feature store is now but not the concept itself, many companies before had feature stores before it was called as such.

[@madhu1987ful]

I didnt exactly understand the importance of feature store. I can as well save these features in a simple delta table and use them later. Why feature store ? We are just saving these columns in a table, but not actually saving the feature definition itself !!

[@paramitasen07]

I hv the same question