It's about performance: building fast web services and benchmarking Spark

Let's meet and talk about Spark and FastAPI! We have two new talks and a shiny new office for you :)

• Level: beginner
• Food and drinks: yes

Agenda:

Part 1:
by Malte Harder (Senior Data Architect @Blue Yonder)

I will give a tour through the FastAPI library. It is advertised as a high performance, easy to learn, fast to code and ready for production framework to build OpenAPI compatible web services. In the talk, I will give a practical introduction to the library and compare it to Flask, which is often used as a Python web framework when building API services.

While it is helpful to have some prior knowledge of building web applications in Python, the talk will only assume some Python experience to be able to follow.

Part 2:
by Sude Saplik (Data Engineer @Blue Yonder)

„Deep Learning Pipelines“ is a framework that provides high-level APIs for buiding scalable deep learning applications in Python using Apache Spark. Its vision is to make deep learning simple. In this talk, I will introduce the framework and give a brief introduction to Spark, PySpark, possible use cases and potential alternatives.
I will first cover the abstract concepts and then continue demonstrating practical examples of this framework. Last but not least, I will show you how I used the framework to study the performance of Spark applications in my master thesis.

I assume that the audience is interested in machine learning and knows some basic Python.