Kramer is the creator of Pipecat (the most widely used open-source voice-agent framework) and CEO/co-founder of Daily, and his talks and writing are dense with hard-won production lessons on getting voice agents from a good demo to >99% reliability — latency budgets, turn-taking, interruption handling, and choosing models that balance intelligence vs. speed. Attend for a candid, build-it-for-real view of the voice AI stack from someone shipping the infrastructure beneath it.

Speaking style

Battle-testedDeeply technicalBig-picture

Recent talks (4)

Pipecat Cloud: Enterprise Voice Agents Built On Open SourceAI Engineer World's Fair 2025 · Jul 2025

Your realtime AI is ngmi — Sean DuBois (OpenAI), Kwindla Kramer (Daily)AI Engineer World's Fair 2025 (Voice Keynote) · Jul 2025

VapiCon 2025: Foundations of Voice AI with Kwindla Kramer, Anoop Dawar, Dr Alberto Montilla & moreVapiCon 2025 · Oct 2025

WebRTC and AI in 2025 — Sean DuBois and KwindlaDaily · 2025

GitHub

@kwindla

Recent writing (6)

Open source, multilingual transcription from NVIDIA: Nemotron 3.5 ASR · blog · Jun 2026
NVIDIA Nemotron 3 Super for Voice AI · blog · Mar 2026
Benchmarking LLMs for Voice Agent Use Cases · blog · Feb 2026
Building Voice Agents with NVIDIA Open Models · blog · Jan 2026
You don't need a WebRTC server for your voice agents · blog · Jul 2025
Advice on Building Voice AI in June 2025 · blog · Jun 2025

Podcasts & interviews (3)

Building Voice AI Agents That Don't Suck with Kwindla Kramer (#739) · The TWIML AI Podcast · Jul 2025
Voice AI is About to Get Loud with Kwindla Kramer · ODSC's AiX Podcast · Sep 2025
Building Production Voice AI That Actually Works: Lessons from Daily & Pipecat Co-Founder · Deployed: The AI Product Podcast · Jul 2025