Case Study: Sync achieves rapid AI video scaling with Modal

How sync. uses Modal to lipsync 100 hours of video a day

sync., a research lab specializing in AI video models for lipsyncing, faced the challenge of scaling its technology from a research prototype into a revenue-generating business. After outgrowing Google Colab and encountering infrastructure headaches with AWS Lambda and slow iteration times with other AI-as-a-service providers, they needed a solution that could handle GPU-intensive workloads and enable rapid deployment.

By implementing Modal's platform, sync. gained the ability to deploy code changes in seconds, iterating up to 95 times a day. This allowed them to scale their operations dramatically, processing over 100 hours of video daily using Modal's parallel execution on GPU containers. The solution from Modal enabled the small team to focus entirely on R&D and model training, shipping 10 major model variants and over 1000 iterations in a year to become a successful business.

Sync

Prady Modukuru

Co-Founder & CEO