Case Study: Quora runs LLM-generated code safely in Poe using Modal Sandboxes
Key results
The challenge
Quora needed to safely execute LLM-generated code within Poe, its AI chatbot platform, with security isolation, low latency, and reliability for millions of users. Building the capability in-house would have taken months of development.
The solution
Quora adopted Modal Sandboxes to run Python code in isolated containers. The team used Modal's built-in security features, scaling infrastructure, and networking primitives rather than developing them internally.
“There would be a lot of edge cases and unknowns if we built code sandboxes ourselves...We offloaded this to Modal and are actively saving 2 engineers' worth of ongoing engineering time.”
HYHwan Seung YeoDirector of Engineering, Quora
The results, in context
Quora reported saving 2 engineers' worth of ongoing infrastructure maintenance time by offloading sandboxing to Modal, and shipped months faster than building in-house. The setup was stress-tested to handle 1,000 Sandboxes per second across Poe's user base.