Behind the VoiceAI: STT & TTS Models Decoded
No telephony. No orchestration. No infra talk. Just STT and TTS the models you're actually choosing between, swapping out, and arguing about in production.
Behind the VoiceAI brings together Voice AI builders founders, engineers, and product teams for an evening focused entirely on the model layer of the stack. This is where most user complaints, latency headaches, and "why did we switch providers" stories actually originate and we're getting into it directly.
What to expect:
🎙️ Panel Discussion - Builders who've shipped with STT/TTS models in production share what worked, what broke, and what they moved away from
💬 Live Q&A - Bring your production questions, get real answers from people who've been there
🥂 Refreshments & Networking - Time before and after to connect with fellow builders
Supported by Speechmatics & Telnyx
Speechmatics brings deep expertise on the STT side accuracy, accent handling, and multilingual recognition.
Telnyx brings the infrastructure lens that connects voice models to real-world telephony. Both are core to how this community builds.
📋 Agenda
5:30 – 6:00 PM Doors open & welcome
6:00 – 7:00 PM Panel Discussion - STT & TTS models in production
7:00 – 7:20 PM Live Q&A
7:20 – 8:30 PM Refreshments & Closing networking
Discussion themes on the panel:
Proprietary vs. open-source models, where each makes sense (cost, control, latency, fine-tuning)
Accuracy & accent handling, especially Indian languages and code-switching
Voice cloning & naturalness, where TTS models are heading
Real-time latency, what it takes to hit sub-second response for live calls
"The model I moved away from" honest takes from builders who've switched providers
📍 Venue: HSR Layout, Bengaluru (exact address to follow)
🗓️ Date: Saturday, 12th September 2026
🕡 Time: 5:30 PM – 8:30 PM
👥 Limited to 50-60 builders
RSVP to secure your spot