How Harvey built their internal cloud agent platform

Harvey, the largest legal AI company, recently published a blog post on how they built Spectre, their internal cloud agent platform. What makes Spectre different is that it's collaborative by design. Engineers, PMs, and researchers all contribute context to the same agent, not parallel ones running on separate laptops.

Join us for a live conversation with Joey Wang, Engineering Lead at Harvey, to go behind the blog post. We'll dig into the architecture decisions, the tradeoffs, what didn't work the first time, and what Harvey is learning about how cloud agents change developer productivity at scale.

What we'll cover:

• Why local coding agents break down once a team needs to share context, and what changes when you move the loop to the cloud
• The core technical primitives behind Spectre: durable runs, ephemeral sandboxes, the harness, and why each call was made
• Live Q&A with Joey on the decisions and lessons that didn't make the blog post

Join this live session for an exclusive look behind the scenes.