S15E1: Building the Infrastructure for Voice AI with Russ d'Sa, Co-Founder and CEO of LiveKit

The Room Podcast's season 15 premiere features Russ d'Sa, co-founder and CEO of LiveKit, a platform for building real-time voice, video, and AI agents. The episode traces d'Sa's career arc from his childhood in Canada and Silicon Valley through roles at 23andMe and Twitter, before founding LiveKit. The conversation covers how LiveKit evolved from an open-source project for live audio/video into infrastructure for AI voice interfaces, particularly after OpenAI adopted it for ChatGPT's voice feature. D'Sa describes this as an "accidental AI company"—the platform wasn't originally designed for AI but found product-market fit there. Key technical topics include why voice AI requires different infrastructure than text-based AI, what still prevents AI voice conversations from feeling natural, and how voice agents differ fundamentally from text agents. The episode also explores LiveKit's scaling journey, including unexpected challenges, investor relationships, and the company's position in the broader AI stack between models and consumer applications. Beyond the business narrative, d'Sa reflects on how fatherhood has reshaped his thinking about success and values. The show notes indicate a structured interview format with 14 distinct topic timestamps, covering both biographical elements (childhood, career transitions, personal influences) and technical/business substance (infrastructure decisions, product roadmap, competitive positioning).

Verdict: A solid founder interview with genuine technical substance and timely relevance to voice AI, but limited evidence of breakthrough insight or exceptional depth beyond the show notes.

Technology rubric

• Technical accuracy 3/5 20%

  Show notes confirm LiveKit's real role in voice AI infrastructure and OpenAI partnership, but lack detail to verify technical claims about voice vs. text agent differences or specific infrastructure challenges.

• Insight beyond the headlines 3/5 20%

  The "accidental AI company" framing and discussion of why voice requires different infrastructure offer some original perspective, but show notes don't reveal whether the episode goes deeper into forward-looking analysis or stays at founder

• Clarity for non-experts 3/5 15%

  Topics like real-time infrastructure, AI agents, and voice interfaces are technical, but the biographical structure and focus on product decisions suggest accessibility; unclear from notes whether jargon is explained or assumed.

• Substance 3/5 15%

  14 timestamped segments suggest structured content covering career, product evolution, and technical challenges; mix of biographical material and business/technical substance, but show notes don't indicate depth or evidence of padding.

• Delivers on its title 4/5 15%

  Title promises "Building the Infrastructure for Voice AI" and the episode clearly covers LiveKit's role, technical differences in voice infrastructure, and roadmap; biographical elements add context but don't dilute the core topic.

• Value for time 3/5 15%

  61-minute runtime with 14 segments suggests reasonable pacing; mix of personal narrative and technical content may appeal broadly, but show notes don't indicate whether the episode maintains momentum or includes slow stretches.

How we score Technology

Worth it if you're building voice AI products, interested in infrastructure-layer startups, or want to understand how real-time communication platforms are becoming AI infrastructure.

Based on the episode description. AI-generated; it can be wrong.