Join us for the Google I/O Extended Saudi 2025, a must-attend developer conference hosted by GDG Cloud Saudi in Riyadh! As part of Google’s renowned global I/O Extended program, this free one-day event will gather approximately 350 top-tier engineers, CTOs, product leads, and tech enthusiasts. 🚀
Dive deep into the latest in AI, Google Cloud, Android, Flutter, web performance, and DevOps through streamed Google I/O keynotes, local tech talks, and hands-on codelabs. Don’t miss the chance to gain practical skills and network with experts and peers in the field. Every attendee will also receive an eco-friendly plant gift, symbolizing our dedication to sustainability. Be part of this unique opportunity to connect, learn, and innovate. Register now to secure your spot and join a vibrant community of continuous learners in Saudi Arabia!
Agenda
1:00 PM: Check-in & Coffee
2:00 PM: Keynotes
Get to know about our Google Developer Group in Saudi Arabia from the leads!
2:15 PM: Event Driven Architecture
Discover how Event-Driven Architecture (EDA) enables scalable, real-time systems by reacting to events as they occur.
2:45 PM: Engineering Resilience at Scale: How HungerStation Handles 345M+ Daily Requests
Discover how HungerStation engineered a hyper-resilient platform capable of managing over 345 million daily requests—without breaking a sweat. From architectural strategies to real-time recovery mechanisms, learn what it takes to deliver seamless performance at extreme scale in one of the region’s busiest platforms.
3:15 PM: Beyond "It Depends": A Practical Guide to Google Cloud Databases
"Which database should I use?" For developers, the most common answer is also the most frustrating: "It depends." This session is designed to take you beyond that vague answer. We'll give you a practical, actionable framework for making concrete database decisions on Google Cloud. We will evaluate the entire data landscape—from relational options like Cloud SQL and AlloyDB, to the globally-scaled Spanner, NoSQL powerhouses like Firestore and Bigtable, and even Memorystore for caching. You'll leave this session armed with a clear decision-making toolkit, ready to confidently choose the right database for your project's specific needs without ever having to say "it depends" again.
3:45 PM: BREAK (Snacks & Coffee)
4:45 PM: Automate your expenses using Firebase
Captures data directly from iPhone messages triggered on the device, then sends it to an AI-powered processing flow—with live updates seamlessly appearing in your app—all powered by Firebase’s real-time ecosystem.
4:45 PM: From Data to Decisions with Google Data & AI
Ever wondered how your favorite apps predict what you’ll like next? That’s the power of data and AI! In this session, we’ll demystify the exciting world of Data Analytics and explore the latest AI breakthroughs from Google. We will show you how these smart tools turn confusing information into clear stories and smarter decisions, impacting everything from business to daily life.
5:15 PM: Memory and Context Engineering in Agentic Systems
This session explores the latest advancements in dealing with memory and context in agentic systems.
5:45 PM: Leveraging Google's ADK and MCP to build powerful AI agents
This talk demonstrates building production-ready AI agents using Google's Agent Development Kit (ADK) through a real-world CV generation system. We'll explore multi-agent architecture patterns, local LLM deployment, and human-in-the-loop workflows that create tailored CVs from job descriptions.
6:15 PM: Serving Vision-Language Models with vLLM
Serving large language models is already hard, but serving multimodal LLMs that understand both language and vision is even harder. In this technical talk, we’ll explore what it takes to deploy and run Vision-Language Models (VLMs) efficiently in real-world applications. You’ll learn the unique challenges of serving VLMs—like image processing, context formatting, and memory overhead—and how vLLM addresses them through innovations like PagedAttention, continuous batching, and support for multimodal inputs. We’ll break down how vLLM enables high-performance, scalable inference for models like LLaVA and Gemini. To bring the concepts to life, we’ll showcase a simple but powerful demo where an AI system narrates scenes from a camera feed. If you're building LLM-powered apps that go beyond text, this talk is for you.
7:00 PM: DINNER (Pizza 🍕)
