Pittsburgh TechFest 2026

Imagine a development environment built on a package repository more than twice the size of Ubuntu's or Fedora's.  One that can be brought up on a new machine simply by cd'ing into your worktree.  Using a single tool with built-in support for the package ecosystems of dozens of programming languages.

Imagine a CI/CD pipeline that uses this same tool to provide caching of build artifacts for everyone on your team.  An OS that provides atomic updates and trivial rollbacks.  And cloud deployments that precisely replicate EC2 images & docker containers.

Come to this talk to learn how Nix provides all of these capabilities, and how it can be adopted incrementally into your projects.

Assembling a team of AI agents shouldn't mean hand-crafting each one from scratch or stitching together agents and skills from popular repos. With Project Squad and the GitHub Copilot CLI, you simply describe your project, requirements and tech stack and Squad casts the right team for the job.


In this live, hands-on session, you'll watch agentic SDLC in action as Squad casts and orchestrates a team of specialized agents: breaking down your requirements, routing each one to the right "team member" based on skillset, and driving the work forward from start to finish, through pull requests, tests, and CI. We'll go from idea to a running app, narrating the decisions, the hand-offs between agents, and the quality gates along the way, compressing what would normally take a disciplined team days or weeks into a single session.


You'll leave with a clear mental model of how agent orchestration across the SDLC actually works, where it shines, where a human stays in the loop, and how to start experimenting with this workflow in your own repos right away.

Like other metaphorical objects that float on the water, Docker has a lot hidden below its surface.  Many of us use Docker or other containerization platforms to run our applications.  But what are containers and images, and how do these two concepts come together in Docker?

In this talk we'll build a simple image and start a running container from it.  Then, like the curious engineers we are, we'll start poking and prodding this container and its image to learn what's actually going on under the hood.

We'll see how Linux namespaces and cgroups are used to lie to the running container process about its networking and filesystem environment and its available resources.  We'll see how the overlay filesystem is used to stack the layers of the image up into what the container sees as a unified file system.  And we'll see how the image is stored as a simple collection of tarballs and json files.

Finally, we'll look at other ways to build images, and see what's different about the resulting images and containers.