This presentation will explore how AI is revolutionizing mobile experiences and the scaling of Facebook while maintaining the quality users expect. We will discuss the evolution of mobile beyond smartphones, with devices like Meta’s Ray-Ban smart glasses, and the role AI plays in making these innovations possible. The presentation will highlight how AI-driven advancements are shaping the future of seamless, scalable, and personalized mobile experiences.

We cover an exciting 2-month journey to integrate Meta AI (Meta's Ai-powered chatbot) into Facebook Feed to provide users the ability to ask the chatbot questions about content in posts on their feed. We dive into challenges around making the experience relevant, interesting, safe and private. We go into details of prompt & response quality - role of multimedia signals in grounding responses as well as role of humans (raters, users, dogfooders) in creating a flywheel to dramatically improve responses. We also cover how we scaled the experience to US+13 countries diving into latency & capacity savings and go into some of trade-offs therein.

In this talk, we'll understand the significance and benefits of Generative AI and learn how we can build Generative AI powered mobile apps with Gemini API and Google AI Studio. This talk will cover examples and cool demos of building content summarization, product recommendation, product decision assistant systems along with code walkthrough, developed during Palm API and Maker suite and Gemini sprints with Google. This talk will also list the benefits of Generative AI for mobile/software developers.

This presentation talks about what it takes from the product and infrastructure perspective to deliver relevant content to the user regularly. In this talk, we will talk about video generally in the industry and at Meta, the key pieces of mobile delivery, video unification and how it has impacted our delivery systems and some new technologies we are able to innovate on due to our new unified product and infra.

We know users expect fast, seamless mobile app experiences. It’s not only important for users, but it’s also a critical business practice. In this talk, we will share how Meta's approach to app quality and deep dive into specific quality domains: Crashes, Correctness, and Video Quality. I look forward to showing you how our methodical approach drives business results, brand perception, and user value.

1. Introduction to Sentinel:
   • Sentinel is an in-house platform built at Uber for real-time monitoring of key business metrics.
   • Automatically monitoring conversion rates over relevant dimensions and providing anomaly alerts to teams.
2. Key Features of Sentinel:
   • Real-time and dynamic alerts.
   • Automatic setup of alerts at any chosen dimension, with city level being the default for up to 300 cities per metric.
   • Unique focus on user impact, offering alerts based on the number of unique users affected and the average number of errors per user (specifically for mobile analytics).
3. Benefits of Using Sentinel:
   • Illustrate how Sentinel can increase the reliability of UI components and overall product features.
   • Real cases where Sentinel detected incidents, sometimes even before they were reported by other systems, showcasing its effectiveness in proactive incident management.
4. Current Achievements and Status:
   • Sentinel has detected over 14+ L4+ incidents uniquely or first.
   • Note the system’s high Signal-to-Noise Ratio (SNR) and its extensive adoption within the teams at Uber.
5. Conclusion and Q&A:
   • Strategic advantages of integrating Sentinel into their systems.
   • Open the floor for questions to address any specific inquiries or concerns from the engineers.

Businesses on WhatsApp seek sophisticated business management tools while retaining the simplicity and familiarity of the WhatsApp app. This talk takes you through how we extended the multi-device protocol to achieve a good balance of sophistication, familiarity and security.

In the last year, Threads was launched and is now a service supporting 200 million monthly users. How did the Threads iOS team maintain performance during this period of incredible growth, which also enabling a high velocity of new feature releases? This talk will detail how the Threads team thinks about performance, the key metrics we monitor to keep the app healthy, and dive into some case studies impacting publish reliability and navigation latency.

A few years ago, Meta embarked on a massive undertaking to translate our entire Android codebase from Java to Kotlin. We developed an automated tool called the "Kotlinator" that has already helped us convert over 10 million of lines of code—including some of the most complex files—while minimizing interference with developers' daily work. Despite having one of the largest Android codebases in the world, we're now well past the halfway point and still going. Join us to learn about the tradeoffs we've made to support automation, seemingly simple transformations that are surprisingly tricky, and how we're collaborating with other companies to capture hundreds more corner cases.

Our kitchens are our playground to display our skills and creativity.

In this talk you’ll visit my humble kitchen to learn my recipes on 10 essential architectural models, design patterns and their roles. You’ll learn how to apply these to your apps just like reading a cookbook from a Chef of a 5 star Michelin restaurant.

Join me as I wear my chef hat and help you understand the most effective ways to use architecture and design patterns in a delicate sense to delight our users.

What it took to rearchitect our mobile network stack to be less than half its size.

We'll be learning about how meta handles the scale of development we do and the unique challenges we face in supporting thousands of developers across multiple apps with interdependencies and millions of lines of code