📄️ Architecture
The Online Feature Store (OnFS) is part of BharatMLStack, designed to support real-time ML workloads through low-latency feature retrieval and flexible feature ingestion pipelines. It ensures that features generated offline or online are immediately accessible for inference.
📄️ Data Formats
In this section we will go through the data-formats which is at the hear of online-feature-store, it's inspired form other storage efficient formats like parquet & arrow, but custom made to deliver in constraint environment. The two key data-formats are:
📄️ Benchmarks
Summary
📄️ Key Functionalities
Overview
📄️ Release Notes
Version 1.0.0 🚀