Presentations
Join us for a live recording of the AskNoah show where we'll be discussing all the things happening around the Ubuntu world!
You need your own data to be useful with AI and large language models (LLMs). How do you go about that? There has been a lot of focus on making applications intelligent using the power of AI with the help of (LLMs). We will cover how to add intelligence to existing applications with an overview of different techniques starting from prompt engineering, using vector stores to improve the results, and building intelligent agents to solve a problem.
You will learn about the techniques using open-source frameworks to build LLM-powered applications like LangChain & LlamaIndex.
Open source is known as the remarkable collaboration of passionate minds coming together to share tech ideas and solutions for the good of the world. Yet ideally the benefit of open source collaboration is a two-way street. Join us as we look at ways to translate what we do in open source into resume-worthy skills.
State Space Models (SSMs) are emerging as a powerful alternative to the transformer-based architectures that dominate today's language models, offering greater efficiency and scalability for modern AI applications. This talk explores how SSMs are transforming AI by reducing computational overhead while maintaining strong performance. Attendees will gain insights into the fundamental principles of state space models, their real-world applications, and the cutting-edge research being conducted at Kwaai to further enhance their capabilities.
This October, I started following along with Linux from Scratch by Gerard Beekmans with the hopes of constructing my own Linux distribution. As with most things computer-related, the simple process I expected turned out to be a tedious, complex, and tiresome one, especially for a relatively inexperienced Linux user like myself. Join me as I tell the story of how I finally made my distribution and try to show the outstanding benefits "doing it yourself" can have with respect to computers and beyond.
Confidential Computing is a technology that enables processing of sensitive data in a secure enclave that is a Hardware based trusted execution environment (TEE). With this technology, data in use becomes inaccessible to even internal administrators, thus enabling technical assurance of data security instead of operational assurance. In this session, we will cover how OpenStack can be used as a cloud control plane to manage a Confidential Computing Infrastructure, its provisioning and life cycle management.
The Zuul project gating system is used to test changes across multiple git repositories and code review systems before they merge. For the past 12 years, it has relied on an external program called Nodepool to manage test resources and image builds.
A brief overview of Zuul for those who are new to the project, and share how to manage test nodes and images with Zuul's newest features.
Recognizing, troubleshooting, and remediating incidents on the services your team owns and runs is hard enough; when the incident is actually happening to an upstream vendor, what can you do? This session will discuss how to plan for vendor incidents, and what to have on hand when someone else is having a very bad day.
Ever wondered how container images are organized? Are you curious why some containers take forever to load and start while other containers start right away? Did you know that "docker build" is not the only way to build container images? Did you know that you can build containers without Docker? Do you want to learn how to build container images without any container tools?
If you answered yes to any of these questions this talk is for you! You'll learn how container images are structured and you'll learn different ways you can build container images.
The ability to provide useful maps is a key feature for many applications, but using commercial services like Google Maps or Mapquest can be expensive and limiting. MapLibre provides a robust and flexible mapping library for both web use (using Typescript) and native mobile operating systems (using C++ and wrappers) and it is completely open source. The MapLibre community is also very active and transparent, insuring that the project will continue and grow in the future,