#machinelearning #featurestore #datatransformation
Data Transformation Layer (Feature Store) is key step in data engineering pipeline that pre-computes critical data elements required for ML model thereby accelerating ML development
Data Transformation rules are fed in from Data Analysis, Domain Understanding, ML Engineering phase
The final resultant repository helps in
Efficient sharing of features across projects
Helps organization build intelligent solution at scale
Easily deploy-able pipelines into production
Reproducible and consistent definitions
You can also check for uber michelangelo platform on how they build scalable pipeline
https://eng.uber.com/michelangelo/
