The best way to share machine learning models is through interactive web applications that let stakeholders try the models for themselves. By playing with the model through a graphical interface, stakeholders can develop an intuition for how the model works in a way that would be impossible by analyzing average performance metrics or hand-picked examples.

Building web applications for machine learning requires significant knowledge of web development (css, js) and web hosting which are not in the typical data scientist's tool box. 

In this talk, we will introduce a gradio, an open source python library for building and sharing machine learning applications entirely in python. 

Isn't dependency injection a Java thing? Not really!

Have you used Pytest? Fixtures are a form of dependency injection. What about just passing in file-like object to `json.load`? Also dependency injection! There's also FastAPI, one of the most popular Python web frameworks now days, that has depedendency injection as one of it's key features.

You see, dependency injection is very widespread in the Python world, it just looks so natual because of lack of a `new` keyword and first-class function support that you might not even know you are using the technique. Nonetheless it can help to be aware of the theory and jargon around it, which is what we'll explore in this talk by way of a joint refactoring exercise.

The Python core developers keep pushing out new functionality every six months.  Python 3.6 just went out of support and Python 3.11 is in pre-release.  This talk will look at the new features promised for 3.11 and maybe provide some opionated views on their usefulness or their potential to make Python code harder to understand.

The world has lost the ability to pip install... what do we do?!?! This talk explores a hypothetical Doomsday Scenario where the Python Package Index has gone offline. We examine three ways we can import libraries we have not installed.

After this talk, attendees will have a better understanding of what happens when a package is imported into Python.

Have you experienced issues with your Python code while wrangling new data sets and crunching big numbers? Have you wondered what are the places where you could make your Python scripts run faster? Let's talk! Today, I will cover a few options to speed up your data processing operations. This talk can be especially valuable for those who are just starting out in Data Science and Python 
development. Nonetheless, everyone interested in efficient Python is welcome!

A recommender system (recsys) is a no-brainer investment for any service that offers users a wide choice of items. This talk will teach you how to design and build one yourself, informed by lessons learnt shipping a recsys to 5 million MAUs. We'll start with the simplest heuristics and ML models that do the trick, focusing significantly on system design. We'll discuss how to compute, store, and serve recommendations - and also how to build a robust data system with long-term value. We can also nerd out about machine learning and idea jam at the end.

I write at bakerwho.github.io and tweet at @bakerwhodata sometimes. HMU

Prefect is air traffic control for the modern data stack. Monitor, coordinate, and orchestrate dataflows between and across your applications. Build pipelines, deploy them anywhere, and configure them remotely. You might just love your workflows again.

Learn how to get up and running with Prefect!

The MicroPython part of Evezor Modular Automation System.

or how to link an input device to a liniar actuator over CAN Bus.

I will maybe spend 5 min talking about my anser to  "What is / Teach me programming?" 

For sure I will spend a few moments talking about Python on FPGAs like Fomu and have a few to give out.

https://micropython.org

https://github.com/adafruit/circuitpython

https://github.com/evezor/Edge_Boards

https://github.com/CarlFK/px/blob/master/For_Teachers.md

https://tomu.im/fomu.html

A walkthrough of the Asynchronous Gateway Server Interface, the API that powers Uvicorn, Starlette and FastAPI.

We'll go through some history, the spec itself, how to write an app and how to write middleware.

By the end of this talk you should have a good idea of what is going on under the hood in your ASGI web server and framework.

This talk explores the relationship between Kubernetes/cloud infrastructure and the most common Python ASGI app deployment topology (Gunicorn with Uvicorn workers). We'll look at the issues that you'll encounter with this deployment strategy and how you can solve them and make your code simpler and more robust by ditching Gunicorn and taking control of your application.