​Session Overview

​Voice agents have crossed the demo stage. The hard part now is making them reliable in production.

​Real users interrupt. Networks fluctuate. Background noise, voicemail, latency, and unpredictable conversations expose failure modes that never appear in a polished demo.

​In this hands-on session with Jesse from LiveKit, we'll build a production-ready voice agent from the ground up while unpacking the engineering decisions behind responsive, reliable conversations. You'll learn how modern voice stacks are designed, where most production failures originate, and how leading teams continuously improve their agents using real-world conversations instead of guesswork.

​Whether you're building customer support agents, AI SDRs, assistants, or internal copilots, you'll leave with practical patterns you can apply immediately.

​What you'll learn

• ​Why voice agents fail in production. The hidden failure modes that only appear with real users, real conversations, and production traffic.

• ​Latency, interruptions, and turn-taking. Understanding where every millisecond goes across STT, LLMs, and TTS—and how to make conversations feel natural.

• ​Pipeline vs. realtime architectures. Choosing the right voice stack based on latency, flexibility, quality, and cost.

• ​Building self-improving voice agents. Capturing production conversations, evaluating failures, and continuously improving prompts, tools, and workflows.

• ​Live implementation. Build a production-ready voice agent with LiveKit and learn practical engineering patterns that scale beyond demos.

​RSVP if you're building voice AI agents!!!

​About Future AGI

​Future AGI is an open-source AI simulation, evaluation, and observability platform. Teams use it to simulate agents before they ship, score them against real failure modes, and keep watching them in production so quality doesn't quietly degrade over time. Self-hostable under Apache 2.0.