Real-time AI

How real-time AI avatars work, where latency hides, and what it takes to make a live conversation feel natural.

Start with why I joined Anam, then follow the engineering: the latency budget, the avatar models and the product decisions around them.

Suggested reading order

Joining Anam as Founding Product Engineer
10 Feb 2025· 3 min read
I'm joining Anam to build out the Lab product and SDK experience, bringing my background as an engineer and founder to real-time AI.
What 100ms means for AI conversation latency
10 Oct 2025· 7 min read
Our testing at Anam pointed to a 200–300ms sweet spot for conversational timing. Here's why that leaves so little room for the engineering.
Shipping Cara 3, Anam's real-time avatar model
17 Feb 2026· 2 min read
What Anam's Cara 3 avatar model does, how it turns audio into a speaking face, and where model generation fits into a live conversation.
Building Anam Lab for cara-4
14 Jul 2026· 4 min read
My work on the Personas home, creation flows and portrait previews for Anam's cara-4 release, helping people build and talk to an avatar of their own.