Uploaded June 2026 | Updated September 2026, 2 weeks ago
Edge AI Voice Tech in Smart Glasses: Latency, Barge-In, Beamforming & NXP's Wearable Platform
This episode of the EdgeVerse Techcast explores how voice technology is adopting edge AI in smart glasses and why glasses are emerging as a natural, always-available interface for conversational assistants without relying on screens. Guest Laurent Pilati of NXP (AI software lead in the ACE organization) explains that real business value comes from integrated AI solutions spanning vision, voice, GenAI, physical AI, and agentic AI. The discussion covers what makes voice interactions feel human—low latency, barge-in, and backchanneling—and the engineering challenges behind them, including audio capture quality, wireless connectivity, inference time, and power constraints. Laurent describes real-world noisy-environment solutions such as fixed and adaptive beamforming, echo cancellation, and speech enhancement, and outlines NXP's software building blocks (Conversa and EAP), i.MX RT700 hardware, and Audio Lab validation, plus advice to start from user experience and leverage proven frameworks.
Episode Resources:
• Laurent Pilati keynote at AWE (youtube.com/watch?v=0Gh3_mFwSdw)
• Essential Audio Processing Library (nxp.com/design/design-center/software/embedded-software/essential-audio-processing-library:ESSENTIAL-AUDIO-PROCESS-LIBRARY)
• Conversa Voice Suite (nxp.com/design/design-center/software/embedded-software/conversa-voice-suite:CONVERSA-VOICE-SUITE?tid=vanconversa)
• i.MX-RT700 Product Details (nxp.com/products/i.MX-RT700)
00:00 Edge AI Intro
01:08 Meet Laurent Pilati
02:21 Why Smart Glasses
03:58 Natural Voice Interaction
05:55 Hearing in Noise
07:32 NXP Platform Stack
09:22 Developer Advice
11:15 Wrap Up and Resources
Edge AI Voice Tech in Smart Glasses: Latency, Barge-In, Beamforming & NXP's Wearable Platform
This episode of the EdgeVerse Techcast explores how voice technology is adopting edge AI in smart glasses and why glasses are emerging as a natural, always-available interface for conversational assistants without relying on screens. Guest Laurent Pilati of NXP (AI software lead in the ACE organization) explains that real business value comes from integrated AI solutions spanning vision, voice, GenAI, physical AI, and agentic AI. The discussion covers what makes voice interactions feel human—low latency, barge-in, and backchanneling—and the engineering challenges behind them, including audio capture quality, wireless connectivity, inference time, and power constraints. Laurent describes real-world noisy-environment solutions such as fixed and adaptive beamforming, echo cancellation, and speech enhancement, and outlines NXP's software building blocks (Conversa and EAP), i.MX RT700 hardware, and Audio Lab validation, plus advice to start from user experience and leverage proven frameworks.
Episode Resources:
• Laurent Pilati keynote at AWE (youtube.com/watch?v=0Gh3_mFwSdw)
• Essential Audio Processing Library (nxp.com/design/design-center/software/embedded-software/essential-audio-processing-library:ESSENTIAL-AUDIO-PROCESS-LIBRARY)
• Conversa Voice Suite (nxp.com/design/design-center/software/embedded-software/conversa-voice-suite:CONVERSA-VOICE-SUITE?tid=vanconversa)
• i.MX-RT700 Product Details (nxp.com/products/i.MX-RT700)
00:00 Edge AI Intro
01:08 Meet Laurent Pilati
02:21 Why Smart Glasses
03:58 Natural Voice Interaction
05:55 Hearing in Noise
07:32 NXP Platform Stack
09:22 Developer Advice
11:15 Wrap Up and Resources
