/Events /Global AI Conference Sydney /Schedule
Global AI Conference Sydney
Monday, November 16
Doors open + Networking
Room 1 ·From Prompt to Pull Request: Building with GitHub Copilot, CLI and Agents
GitHub Copilot has evolved from code completion into an AI-powered development platform spanning the IDE, command line, code review, agents and application workflows. In this practical, demo-led session, Saanvi Yerawar and Jimmy Gao will show how developers can use GitHub Copilot to move from idea to working software faster while keeping humans in control.
We will walk through real-world developer scenarios: understanding an unfamiliar codebase, using Copilot Chat to design and explain changes, driving terminal workflows with GitHub Copilot CLI, generating and refining implementation changes, and using Copilot capabilities across GitHub to improve pull requests and team productivity.
Attendees will leave with a clear view of Copilot capabilities they may not be using yet, patterns for applying AI safely in day-to-day engineering, and practical tips for adopting Copilot in teams without turning the session into a sales pitch.
Saanvi Yerawar is a Specialist in App Innovation at Microsoft, working with customers and developer teams on modern application development and AI-powered engineering practices. She focuses on helping teams apply tools such as GitHub Copilot in practical, responsible ways to improve developer productivity, accelerate delivery and modernize software engineering workflows.
I'm a designer who can't code, and I'm shipping apps anyway
I've spent 10 years as a product designer and never learned how to code. But I've always been around products, working out what to solve and why. What I couldn't do was code and build the thing myself.
Now I can, so I'm building Grocero, a grocery price comparison app for Australian supermarkets. There's a frontend for iOS and Android, a backend with scrapers, an API, and a landing page to make the marketing side work. Claude Code wrote all of it while I was orchestrating.
I'll walk you through the workflow that got it there. It starts with the early idea in my head and a lot of chats with Claude the sceptic, and it ends with the guardrails that keep the code safe: a rules file Claude can't break, plan-before-code, and a full code review before anything ships.
Spoiler alert: building turned out to be the easy part. Judgement is the hard part, and I'll show you how I make it work.
Staff Product Designer at Propeller, working on geospatial software for the construction industry. Outside of work I build and ship real apps with Claude Code, without writing the code myself. Nudget, a job application tracker, is live on the App Store. Grocero, a grocery price comparison app for Australian supermarkets, is what I'm building now.
The Engineer in the Loop: Building with Codex
Jeremy Yee shares practical examples and lessons from building software with Codex. He walks through how he gives the agent useful context, guides changes and checks the results. Developers will take away practices they can adapt to their own projects.
Jeremy Yee is a software engineer and Codex Ambassador, and is building Studio, software for planning and coordinating events. He co-hosts Sydney's Codex community events and is interested in how developers can get more useful work out of coding agents through clear context, meaningful feedback and engineering judgment.
Grok Bot at home and at work
Josh explains how he uses Grok Bot at home and at work, from creating fun AI films to handling daily chores like meal planning and negotiating electricity and phone bills.
We'll unpack what Grok Bot is, the design principles behind the scenes, and how you might think differently about your computer after you meet your Bot.
Joshua Crowley is a design engineer at Envato and a SpaceXAI regional lead ambassador. He also runs training and workshops through Jagged Peaks Club. Josh has spent over a decade working in UX and design engineering across government, education and media.
AI Tokenomics Explained
AI systems rely on tokens for every response, but the cost and supply of those tokens can quickly become difficult to manage.
This session explains how token utility, demand forecasting, supply, inference efficiency, and monetization fit together, so builders can make better decisions about capacity, cost, and scale.
Dr. Haritha Thilakarathne is a Senior Developer Relations Manager at NVIDIA. He works with AI builders across the ANZ region to make the most of accelerated computing technologies. Haritha collaborates with developer communities and enterprises to design and deploy secure, and scalable AI solutions.