The Feldera Blog

An unbounded stream of technical articles from the Feldera team

RocksDB: Not A Good Choice for a High-Performance Streaming Platform

We go over our findings when trying to use RocksDB as our storage engine

Rolling Aggregates: the Most Important Feature in Your Real-Time Fraud Detection Pipeline

We explore rolling aggregation--a critical tool in the real-time fraud detection arsenal

Optimizing Kafka input performance

What can we do to our code and our data to maximize Kafka input performance?

Your Streaming Analytics Use Case Needs More State Than You Think

There are two types of streaming users: those who believe their workload fits in memory and those who already know it won't.

Real-time fraud detection using Feldera and Hopsworks

Implementing a real-time fraud detection pipeline using Feldera and the Hopsworks feature store.

Let's make streaming analytics boring

How can we make streaming analytics as user-friendly as batch analytics? Working back from the ideal user experience, we identify the issues with current streaming tools, and explore ways to address them.

Real-time feature engineering with Feldera. Part 2.

We put Feldera to work to implement real-time feature engineering for credit card fraud detection on top of a Delta Lake.

What do we mean when we say real-time SQL?

Feldera can run any batch SQL query, but in streaming mode. We illustrate using a SecOps (security operations) demo.

Incremental Computation, a bad case of déjà vu

Incremental Computation, a bad case of déjà vu