Build a Bilingual Multi-Agent Voice AI System with Vapi

​About Event

​Most voice AI demos work in isolation, but break down when real-world complexity comes in - routing, context switching, and handling multiple languages in the same interaction.

​This hands-on workshop is focused on what it actually takes to build production-ready voice AI systems, designed specifically for senior AI/ML engineers, backend and data engineers, and technical builders working on voice and agent systems.

​You’ll build a live inbound phone system powered by multiple AI agents working together as a coordinated system. A triage agent detects language (English or Spanish) and routes calls in real time to specialized agents handling maintenance requests or answering FAQs from a bilingual knowledge base.

​By the end of the session, you’ll have a working, callable system with real-time agent handoffs, external API integrations, and a setup that reflects how voice AI systems are built in practice.

​This is a practical, builder-focused session designed for engineers and teams working on real systems, not demos.

​WHAT YOU’LL BUILD

​By the end of the workshop, you’ll have:

• ​A live, callable inbound phone system

• ​A multi-agent setup with three coordinated agents:

  • ​Triage Agent (language detection and routing)

  • ​Maintenance Agent (collects repair details and logs tickets)

  • ​FAQ Agent (answers from a bilingual knowledge base)

• ​Real-time agent-to-agent handoffs during live calls

• ​Bilingual voice support (English and Spanish)

• ​Integration with external APIs for dynamic workflows

​TOPICS WE’LL EXPLORE

​Some of the key themes throughout the session:

• ​Multi-agent orchestration and routing logic

• ​Building voice AI systems beyond single-agent setups

• ​Real-time agent handoffs and shared context

• ​Tool calling and API integration in voice workflows

• ​Prompt design and guardrails for voice agents

• ​Multilingual voice systems (transcription + TTS)

• ​Designing systems that handle real-world, non-linear conversations

​AGENDA

​5:30 – 5:45 PM | Doors Open — Arrival & Setup
Check-in, networking, and environment setup
5:45 – 6:00 PM | Introduction & System Overview
Overview of production voice AI systems and workshop walkthrough
6:00 – 6:45 PM | Hands-on Build — Core System
Set up agents, routing, and multilingual voice stack
6:45 – 6:55 PM | Break
6:55 – 7:40 PM | Hands-on Build — Advanced Workflows
Agent handoffs, API integrations, and bilingual flows
7:40 – 7:50 PM | Testing & Live Demo
Run live calls, debug flows, and refine system behavior
7:50 – 8:15 PM | Wrap-up & Next Steps
Key takeaways, extensions, and Q&A

​WHO SHOULD ATTEND

​This workshop is designed for:

• ​Senior AI / ML engineers

• ​Backend and data engineers

• ​Engineers building voice, agent, or RAG systems

• ​Technical product builders

​We prioritize attendees from larger organizations (500+ employees), including technical decision-makers, engineering leaders, and teams actively building or evaluating voice AI systems.

​PREREQUISITES

​To get the most out of the session, you should have:

• ​Basic familiarity with APIs and developer workflows

• ​Comfort using the terminal and editing code

• ​Node.js, Git, and a code editor installed

​You’ll also need accounts for Vapi, GitHub, and Val Town (details will be shared ahead of the event).

​No prior voice AI experience required.



Seats are limited