The GitHub Universe 2026 schedule simply dropped, and it’s stuffed with thrilling classes, demos, and panels masking the potential of AI-powered growth.
In the event you haven’t registered but, right here’s what it’s essential to know. This two-day occasion brings collectively among the biggest minds in tech, with consultants from corporations like AMD, Figma, NVIDIA, Coinbase, Anthropic, and OpenAI main our classes. They’re masking all the things from delegating actual work to Copilot and measuring AI at enterprise scale to fine-grained safety for MCP servers. You’ll even have the possibility to talk with the GitHub workforce one-on-one, get your questions answered, and even decide up some profession recommendation.
Did we point out the Ship & Inform classes the place groups present what they’ve constructed, and companion cubicles the place you’ll be able to demo the most recent tech?
When: October 28-29
The place: Fort Mason Heart, San Francisco, CA
One factor first: register earlier than August 19 and save $300 with Early Fowl passes. Costs go up after that, so if Universe is already in your checklist, now’s the second. The most effective half? You possibly can stack financial savings with our group reductions.
Right here’s a sneak peek of among the classes we’ve got deliberate. You possibly can bounce to the total agenda proper right here. You’ll want to mark your favorites to construct your personal private calendar.
Discover your circulation
Among the greatest moments at Universe occur heads-down: working by an actual downside, configuring one thing by yourself machine, and strolling out with a venture you’ll be able to really use. This 12 months’s catalog leans into that with learnings you’ll be able to take straight again to your repositories.
A couple of classes to start out with:
Cease prompting, begin delegating: Configure Copilot to personal the workKen Muse, GitHub; Mickey Gousset, GitHubLearn how the best Copilot configuration turns AI into one thing you’ll be able to belief with advanced duties. Layer Copilot’s full stack onto a TypeScript app, and also you’ll go away with a working venture and a transparent sense of which functionality suits which job, so you’ll be able to delegate extra and immediate much less.
Inside GitHub Copilot’s coding harness: Optimizing throughout each modelJulia Kasper, MicrosoftShipping a coding agent that works throughout OpenAI, Claude, Gemini, and no matter drops subsequent week takes a harness. See the analysis framework the GitHub Copilot workforce makes use of to check and optimize its agent throughout each mannequin: reproducible benchmarks, 1000’s of autonomous coding duties, and LLM-graded assertions that catch regressions earlier than customers do.
Cease ready by yourself pull requests: GitHub stacked pull requests in practiceSameen Karim, GitHubStacked pull requests allow you to break up modifications into smaller, dependent pull requests that transfer by assessment effectively whereas preserving the total image. This demo builds a stack with the GitHub CLI, opinions it on github.com, and merges every pull request because it’s prepared—so you permit with a workflow you should use tomorrow.
Discover your individuals
Hallway conversations, a query that reframes your entire method, the engineer who already solved the factor you’re caught on. This 12 months’s agenda is crammed with classes for precisely that. Plus, hallway tracks, companion cubicles, and Ship & Inform classes the place groups present what they’ve really constructed.
A couple of classes price your time:
Constructing AI fluency at UPSJared Hatfield, UPSGetting builders to attempt GitHub Copilot is straightforward; getting them fluent with it—utilizing brokers to plan, write, and ship actual work—is the tougher problem. See how UPS strikes builders from consciousness to fluency, why motivation and measurement matter as a lot because the tooling, and the place to focus first at enterprise scale.
Code is the straightforward half: Constructing Residence Assistant within the openFranck Nijhof, Open Residence Basis“Constructing within the open” often means one factor: code on GitHub. However the arduous work begins lengthy earlier than code—concepts, UX, design, structure, the roadmap itself. At Residence Assistant, each step occurs within the open throughout 20,000 contributors and a dozen GitHub organizations. See how the Open Residence Basis runs its entire roadmap with points, tasks, and discussions, together with the tougher components, like being improper in public, fixing it in public, and proving you don’t need to be technical to contribute.
I made my Octolamp suppose with GitHub Copilot CLI hooksBeatris Mendez Gandica, Nuevo FoundationWhat in case your desk lamp may present when GitHub Copilot CLI is pondering? Utilizing the native hooks system, Beatris Mendez Gandica made hers breathe inexperienced when the agent works, go white when idle, and blink purple on errors. No prompting methods, simply system-level lifecycle occasions driving a bodily mild through the WLED API. On this session, you’ll see how Copilot CLI hooks work beneath the hood, and go away realizing tips on how to write your personal for any use case.
Construct what’s subsequent
Wish to know what’s on the horizon? These are the talks that pull again the curtain on the place AI-assisted growth is heading.
The view from the labs: What’s subsequent for AI-assisted developmentCara Phillips, Anthropic; Rohan Varma, OpenAI; Kate Catlin, GitHubThe individuals constructing frontier AI fashions see the place capabilities are heading earlier than anybody else. This interactive panel brings leaders from Anthropic, OpenAI, and different labs that energy GitHub Copilot collectively for a candid take a look at the following two years of AI-assisted growth, and what it means for you.
Open pull requests, don’t merge them: Advantageous-grained authorization for hosted MCP serversNick Taylor, PomeriumHosted MCP servers hand each agent all the things its human can do: OAuth in, broad scope out, one international toggle. However what if an agent ought to open a pull request and go away the merge to a reviewer? See a sample that works immediately: an identity-aware proxy that provides per-identity authorization in entrance of any hosted MCP server with no modifications upstream, demonstrated reside with Copilot doing precisely that.
From writing code to managing brokers: Scaling 50+ companies at GitHubAnjuan Simmons, GitHubGitHub’s Lifecycle workforce traded hand-coding fixes throughout 50+ companies for an agentic pipeline the place AI brokers classify points, analysis codebases, write implementation plans, and open draft pull requests routinely. Get the actual classes and numbers from operating it at GitHub’s personal scale: the way it was constructed with GitHub Actions and Copilot, the place automation pays off most, and why engineers nonetheless intentionally write some code by hand.
Forged your vote!
Three breakout classes are heading to Universe this 12 months, however we want your assist deciding which one makes it to the primary stage. Learn by the descriptions, then solid your vote earlier than August 21. Whichever session wins will even be out there on demand, so even if you happen to can’t be part of us in particular person, your vote shapes what you’ll get to look at later.
Session 1: Why sketching in code issues extra within the age of AIQianqian Ye, Processing Basis / College of Southern California; John Maeda, GitHubAI can generate clear, working code from a brief immediate, so the actual query now could be tips on how to perceive what’s been made. Qianqian Ye attracts on instructing inventive coding at USC and main p5.js to point out why sketching and pondering in code builds the instinct to form techniques, not simply immediate them.
Session 2: Nonetheless open: How AI reshaped open sourceAngie Jones, Agentic AI Basis; David A. Wheeler, The Linux Basis; Priya Pahwa, Django Software program FoundationOpen supply is being stress examined. Whereas AI has accelerated contributions, it’s additionally raised arduous questions on high quality, governance, and the facility of group when bots can contribute at machine scale. On this session, you’ll study why AI makes open supply extra important than ever.
Session 3: The human aspect of AI: How constructing group drove 94% Copilot adoptionBrittany Istenes, IndependentWhat if scaling AI isn’t about higher fashions, however higher human connection? A big U.S.-based monetary companies agency hit 94% Copilot adoption and 78% agent-written code, and the driving drive was group, not tooling. On this session, you’ll uncover how grassroots engagement, storytelling, and innersource made AI one thing builders really needed to make use of and construct collectively.
Study extra
Wish to get much more out of your GitHub Universe expertise?
If you register, you’ll be able to add a Day of Studying cross for $50, which incorporates entry to a full day of studying at GitHub HQ plus one GitHub Certification examination voucher.
Have already got your agenda deliberate? You can too buy a GitHub Certification examination voucher as a standalone add-on for $30 and validate your abilities after the occasion.
See you in SF!
Written by
