How PayPal uses inference graphs for MLOps

Are machine learning inference pipelines getting more complex?
Are you using multiple models for making your predictions?
A streamlined model having patterns such as chaining, fanout or ensemble can significantly reduce the effort for ML engineers and make ML systems more agile.
Sharmili Srinivasan and Saurav Raj – ML engineers – explain the use case of Inference Graphs at PayPal.
This talk was featured at The Fifth Elephant MLOps conference. Watch videos about MLOps practice at Udaan and other prominent organizations at https://hasgeek.com/fifthelephant/mlops-2022/videos