List of videos

Full title: Sponsor Presentations: Blockchain now speaks Python: How to use it on Algorand to build real solutions (Sponsor: Algorand) Presented by: Joe Polny Bring your existing Python skills, existing testing framework, and favorite IDE to start building on the blockchain! It has never been easier for you as a Python dev to add another powerful tool to your toolbox. Why build on blockchain? It offers permanent, transparent record-keeping, enables traceability and provenance, and unlocks opportunities for innovation across industries. All you need is your laptop, a willingness to code, and the same Python toolchain you're already familiar with. We have everything you need to code, including Poetry, Pytest, and Docker, wrapped up in a neat package. To get you started quickly, some basic templates are included too! The best part? You can code using the free and open-source GitHub Codespaces virtual IDE, so no need to install anything locally. Within the hour you'll deploy fundamental blockchain primitives, including data storage CRUD, getter/setter methods, logic evaluations, and role-based access control. Be the first in your team to know blockchain.

Full title: Sponsor Presentations: Avoid the top 5 web data pitfalls when developing AI models (Sponsor: Bright Data) Presented by: Jakub Glodek Data Bias: ensuring that the training data is not biased. Biased data can lead to AI models that are unfair or discriminatory. For example, if a dataset for facial recognition software predominantly contains images of people from certain ethnic groups, the model may perform poorly on faces from underrepresented groups. Insufficient Data Variety: AI models require diverse data to understand different scenarios and variations. If the training data is too homogeneous or lacks variety, the model might not perform well in real-world, diverse conditions. Overfitting and Underfitting: Overfitting occurs when a model is too complex and learns to fit the training data so closely that it fails to generalize to new data. Underfitting happens when the model is too simple to capture the underlying patterns in the data. Poor Data Quality: If the training data is full of errors, inconsistencies, or is poorly labeled, the AI model will likely inherit these flaws. Ensuring high data quality is essential for developing reliable and accurate AI models. Ignoring Data Drift: Over time, the real-world data that an AI model encounters may change or 'drift' from the data on which it was trained. This can happen due to evolving trends, behaviors, or environments. Failing to monitor and adapt to these changes can render an AI model less effective or even obsolete. Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/155/2024-05-15T23%3A29%3A58.171807/Bright_Data_ScrapeCon.pptx

Full title: Sponsor Presentations: The open book of Snowflake's AI research: A look inside our generative AI innovations (Sponsor: Streamlit) Presented by: Yusuf Ozuysal As AI becomes more crucial in our daily lives, transparency in AI models is vital. However, many Large Language Model (LLM) systems keep crucial information to train large language models at scale proprietary, contradicting the principles of openness that define the Python ecosystem. This session invites you to explore how Snowflake's AI research team implements openness in building and sharing generative AI developments, prioritizing simplicity and ease-of-use while maintaining efficiency at scale. In this talk, you’ll learn: - How open tools, practices, datasets, and recipes contribute to and inform our research - The recipes we used – including what worked and what didn’t – when training LLMs - Developer-first example use cases of prompting, retrieval, and presentation using the datasets, embeddings, and model trainings we developed Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/161/2024-05-15T23%3A47%3A39.942213/PyCon_Arctic.pdf

Full title: Sponsor Presentations: No Data? No Problem: Zero-Data Model Training with Foundational Models (Sponsor: Covalent) Presented by: Santosh Kumar Radha Ara Ghukasyan In this talk, we introduce a streamlined method for training smaller, task-specific models without the need for new data by utilizing foundational models. We'll cover the process of leveraging existing open-source foundational large models to impart knowledge to more compact models tailored for specific tasks. The focus will be on how developers can apply these techniques using open-source tools to enhance their projects.

Full title: Sponsor Presentations: Finding the needle: a deep dive into the rewriting of Haystack (Sponsor: Haystack) Presented by: Tuana Celik Massimiliano Pippi Haystack is an open-source framework. With Haystack you can compose various NLP tools to build applications, with a particular focus on Large Language Models. Haystack was built before the “ChatGPT revolution”. Same as many others in this industry, we had to question all the existing assumptions in order to adapt, and we had to do it fast. In this talk, we'll explore the motivations behind the refactoring, the challenges we faced, and the outcomes achieved through this intensive process. From rethinking many of the original abstractions, all the way up to growing a vibrant community of users and contributors, we’ll share the key strategies and techniques employed during this journey. Whether you're a seasoned open-source contributor or a curious enthusiast, this talk promises to uncover valuable insights and lessons learned from the evolution of Haystack.

Sponsor Presentations - Rethinking How We're Linking: What to do when speeding things up slows them down (Sponsor: Meta) Presented by: Loren Arthur The common belief that re-writing Modules as CExtensions will improve performance only works to a point. At Meta, we operate at a massive scale with applications that can include over 20,000 shared libraries. When loading that many files, import performance slows down significantly. In this session, I will share how we greatly improved performance by statically linking native extensions into the runtime. Get ready for a quick overview of linkers and loaders, a brief foray into binary layout, and a deep dive into how Native extensions work in Python. Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/152/2024-05-16T03%3A14%3A19.129398/LorenArthurPycon2024.pdf

Sponsor Presentations: Navigating the AI Labyrinth in Higher Education (Sponsor: Western Governors University) Presented by: Jared Plumb Eric Lagally The rapid evolution of Artificial Intelligence (AI) presents unique challenges and opportunities for higher education. This presentation explores the intricate journey of teaching AI within a landscape where its definition and applications are continually expanding. As AI technology advances, academic institutions must adapt to teach a subject that is inherently multifaceted and ever-changing. The session, led by professionals from Western Governors University, an institution renowned for its online and competency-based education model, will delve into the complexities of integrating AI education in higher education. We'll examine the varied interpretations of AI across disciplines, discuss the regulatory challenges universities face in managing AI usage, and share insights on fostering an adaptable AI curriculum that prepares students for a future where AI's role is both transformative and omnipresent. This discussion aims to spark dialogue on effective strategies for navigating the AI labyrinth in higher education, ensuring that students are not only proficient in current AI technologies but also prepared for the ethical and practical challenges of tomorrow's AI innovations. Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/154/2024-05-16T02%3A53%3A09.935152/AI_in_Higher_Education.pptx

Sponsor Presentations - An Introduction to Simulation And Ansys For Developers (Sponsor: Ansys) Presented by: James Derrick Numerical simulation can be complicated and scary. Often, it is the exclusive realm of engineers and the highly trained, but it doesn't have to be that way! Simulation has a lot to offer to anyone interested in investigating the world around them and while the basics may seem overwhelming at first, it really can be valuable, and even fun to have a working knowledge of the steps involved. This presentation is designed to prepare you with a basic grounding in simulation and how it works so that you too can go away and build your own simulations, or more realistically, use simulation tools (such as Ansys) to model every day things. In particular this presentation will take you through an example script of PyMAPDL (our Open Source Python interface for Ansys' original product: Mechanical APDL), modelling a simple 2D bridge scenario. Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/168/2024-05-15T03%3A41%3A40.067703/Intro2Simulation4Developers.pdf

Presented by: Seth Michael Larson Michael Winser Alpha-Omega is investing in the security of open source software ecosystems like Python by staffing security champions focused on improving software supply chain security. In this session attendees will hear from the inaugural PSF Security Developer-in-Residence about improvements that have been made to Python and how through collaboration and knowledge sharing the tides are raised for all ecosystems and users. Slides: https://pycon-assets.s3.amazonaws.com/2024/media/presentation_slides/148/2024-05-16T02%3A31%3A53.213154/PyCon_US_2024_State_of_Python_Supply__XmbR47L.pdf