Kwindla Hultman Kramer is the co-founder and chief executive of Daily, the real-time communications company behind Pipecat, an open-source framework for conversational voice and multimodal applications. His career spans early web infrastructure, gesture-controlled spatial computing, and the real-time systems that make spoken interactions with AI practical.
Kramer began programming professionally during the early commercial web, building web servers, search systems, and publishing tools. He became the founding chief technology officer of AllAfrica, helping develop infrastructure for the African news organization.
In 2006, he co-founded Oblong Industries with John Underkoffler to commercialize spatial-interface concepts associated with Minority Report. As chief executive, he helped develop the g-speak Spatial Operating Environment, which combined gesture, multiple screens, and shared workspaces.
At Daily, Kramer extended that interest in unconventional interfaces into developer infrastructure for live audio, video, and AI. By 2023, he was building applications that combined speech recognition, language models, and synthesized voice. Daily and its open-source community developed Pipecat to orchestrate the audio pipelines, conversational context, model calls, interruptions, and provider integrations behind responsive voice applications.
- Voice is an end-to-end systems problem. Useful conversational software depends on transport, transcription, inference, speech generation, and application state working together under tight latency constraints. Kramer emphasizes production evaluation and instruction-following accuracy over demonstrations that merely sound convincing.
- Turn detection moves down the stack. Kramer initially handled conversational timing in application code, then incorporated it into Pipecat; real-time model APIs increasingly absorb the same responsibility. His work with Google DeepMind on real-time voice workflows illustrates how interruptions, function calls, and changing interface state remain difficult even as the underlying infrastructure improves.
- AI-native software extends beyond chat. Kramer treats today’s agents as foundational components for richer applications combining shared context, background subagents, dynamically generated interfaces, and conversational control. His Gradient Bang multiplayer game explores those patterns, while his local macOS voice-agent experiments investigate conversational systems running on personal computers.
