Thu, Mar 12 · 5:45 PM EDT
Agenda:
5 : 30 - 6 : 00 PM - Check in & Networking
6 : 00 - 6 : 45 PM - Speaker 1
7: 00 - 7 : 30 PM - Speaker 2
7:45 - 8: 00 - Closing
Speaker 1: Snow Zhao
Bio
Snow Zhao
Founder & CEO of AIgentic | Ex-Meta engineering leader. Building the future of professional services with precision & trust.
Snow has decades of Silicon Valley experience building and scaling technology organizations at Facebook, Block (Cash App), and Groupon. She is currently building at AIgentic, an applied AI company designing robust, trustworthy AI systems for professional industries. She leads global teams that delivered secure, high-performance systems used by billions, from product infrastructure behind Messenger’s rewrite and end-to-end encryption, to architecting enterprise AI platforms that bridge traditional sectors with digital transformation. Her work has spanned mobile, security, open-source, federated architecture, AI-powered automation, cross-platform software ecosystems, and more — always guided by the principle that engineering excellence is a human discipline as much as a technical one.
Title
Break the Cloud Lock-in: Building Adaptive Intelligence Orchestration with DIO
Summary
AIaas are often stuck in a binary choice: pay the "token tax" for convenient Cloud APIs or manage the complexity of local open-source models (Llama/Mistral) yourself. The future belongs to a hybrid architecture—but the routing logic shouldn't be spaghetti code.
In this session, we introduce DIO (Distributed Intelligence Orchestration), a new open-source framework designed to bridge on-premise infrastructure and cloud capabilities. We will move beyond theory and look at the code behind the Federated Decision Engine (FDE)—the algorithm that allows you to dynamically route prompts based on privacy, cost, complexity, and more.
Session Takeaways:
The DIO Architecture: A technical deep dive into the open-source framework that decouples your application from specific model providers.
Live Integration: How to implement the DIO SDK to build a "Smart Router" that connects the best of both world for Cloud & On Prem Intelligence.
Speaker 2:
Kiran Rao, Mobile app developer at Somnox.
"Two decades as a developer, got into Android in 2010. Adopted Kotlin in 2017, had my first attempt at KMP in 2019 (when it was still called KMM).
Worked on and off with iOS and web client technologies throughout my career, also dabbled in other cross platform technologies. Full-fledged Mobile developer since 2023."
Title
Leaning on the SQL: Crafting Resilient Data Layers
Description
Most apps need to persist data. The mechanisms for persisting vary, from key-value pairs to relational databases to files and more. SQLite is the de-facto standard for persistence in mobile apps: it ships with every Android and iOS app.
This session looks into best practices for SQLite data persistence, drawing from practical, digestible, real world examples. We look into when one should lean on the SQL versus when to perform data manipulation in application code. We look at some examples of constraints that you might have written in Kotlin/Swift, and how expressing them in SQL instead can offer data consistency and integrity benefits.
A major focus of the session is on migrations. Data schemas inevitably evolve, and executing migrations carefully is essential, as mistakes can have severe consequences, including data corruption. We look into best practices for migration, including when to do migration in application code versus in SQL. We also examine some real world migration examples beyond the simple case of adding a column with a default value.
Finally we look at migrations that cross data sources. This could occur if for example you were persisting data in a key-value store but you now want to migrate that to a database table. This must be done in application code but can you still take advantage of database migrations? Can you cover your migration code with unit tests?