Leads engineering for Runway’s generative video models and the infrastructure that serves them to millions of users.
Session
Main Stage
Date
Tuesday, December 1
Level
Intermediate
// AI Engineering Skill /
Machine learning foundations
// 01 / Bio
Kamil Sindi is the Chief Technology Officer of Runway, where he leads engineering for the company’s generative video models and the real-time infrastructure that serves them. He has spoken about the engineering behind WorldPrompt in Runway’s GWM Worlds 2 research preview, including how long-running generations manage context without exhausting GPU memory, and how to evaluate interactive worlds as they grow more complex.
Before Runway, Kamil was Head of Engineering at Lithic, the card-issuing API platform, and an engineering leader at RapidSOS, the emergency response data platform. He is an active open-source contributor: his managers-playbook repository on engineering management has more than 5,000 stars on GitHub, and he contributes to avtensor, Runway’s Rust library for decoding media directly into PyTorch tensors. He also sits on the board of EliseAI.
// 02 / The session
What It Takes to Train and Serve Video Foundation Models
Abstract
How Runway trains large-scale foundation models for video generation, and how the team scales its infrastructure to deliver those models to millions of customers.
What you’ll learn
How Runway trains large-scale video foundation models
What changes when generative models have to run in real time
Scaling inference infrastructure to millions of users
Skills you’ll practice
AI DEV · BY DEEPLEARNING.AI
Join us at AI Dev
Meet the engineers shipping AI to production, and leave ready to do it yourself.