In the rapidly evolving landscape of AI/ML, the shift from batch to real-time data processing is significant. It impacts how quickly and dynamically we can learn from data, leading to more responsive AI applications. In this session we will explore the shift from batch analytics to real-time decision-making for AI/ML use cases.

We will start by fostering a broad understanding of how to train and score a real-time model on streaming data, facilitated by Kafka/Redpanda, in contrast with traditional batch-processing methods. We will discuss the important aspects of time series data and time-aware features in the context of real-time analytics. Additionally, we will cover how to merge multiple data streams for more complex feature creation and scoring.

This session will include a demonstration using flight and weather data. We'll apply real-time streams to anticipate future air traffic, illustrating a simple application of these concepts that attendees can extend to their own use cases. We will also explore the implications for MLOps in a streaming environment. We'll discuss the adjustments required for real-time data handling and strategies to address the issue of missing data in a real-time setup and how to make decisions when parts of the data streams fail.