Natan Mish - Event Driven Machine Learning | PyData London 2023

Discover how event-driven machine learning overcomes traditional limitations in rapidly changing environments with Apache Kafka, Delta Lake, and Streamlit.

Key takeaways

The traditional way of implementing machine learning models is not suitable for today’s fast-paced and rapidly changing environments.
Event-driven architecture is a solution that can be used to overcome the limitations of traditional machine learning approaches.
The event-driven architecture is a decentralized system where data is processed in a real-time manner and where the model is constantly being updated.
The concept of concept drift is essential in event-driven machine learning, as the relationship between the features and the target variable changes over time.
Online machine learning is a key aspect of event-driven machine learning, as it allows for real-time updates of the model.
Apache Kafka is an important open-source project that enables event-driven architecture.
Commercial offerings such as AWS Lambda also enable event-driven architecture.
Delta Lake is an open-source framework that allows for efficient storage and processing of data.
Streamlit is a tool that can be used to create interactive applications for event-driven machine learning.
The importance of testing and simulation in event-driven systems cannot be overstated.
The traditional monolithic architecture is not suitable for modern fast-paced environments.
Event-driven architecture offers many benefits, including reduced latency, increased scalability, and improved maintainability.
It is essential to consider costs and resources when implementing event-driven architecture.
There are many open-source projects that can be used to implement event-driven architecture, such as Knative.
Testing and simulation are essential in event-driven systems to ensure that they are robust and reliable.

Natan Mish - Event Driven Machine Learning | PyData London 2023

More talks