Snorkel uses Modal for agent simulations
The story so far
Snorkel uses Modal to run millions of agent simulations monthly, leveraging GPU and CPU sandboxes for tasks like evaluation and testing. Modal’s features, such as memory bursting and shared/dedicated endpoints, help Snorkel reduce costs by 20% and improve reliability while allowing quick model deployment. Customers are charged based on the higher of used or requested resources, with ~10-15% of Snorkel’s sandboxes currently using bursting. Snorkel also benefits from Modal’s infrastructure-free model loading and inference serving.
AI-written from the reports · Updated 12 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- Modal BlogHow Snorkel runs realistic agent simulations with Modal
Snorkel uses Modal's GPU and CPU sandboxes to run millions of agent simulations per month, supporting various task evaluations. Through Modal's memory burst feature, Snorkel reduces sandbox costs by 20% and improves runtime reliability. Modal's serverless GPUs and shared/dedicated inference endpoints enable Snorkel to quickly deploy and test custom models.