Dylan Fox, Founder and CEO of AssemblyAI, built one of the leading speech AI infrastructure companies while starting out as a solo founder. Based in New York, Fox studied Business Administration with a focus on Business Economics and Public Policy at George Washington University, and worked as a research engineer at Cisco on early AI products before striking out on his own.
Fox founded AssemblyAI in 2017, inspired by the emergence of the first generation of voice-controlled assistants, such as the Amazon Echo. At the time, most speech technology was built either for consumers or large enterprises, leaving a gap for developers who needed reliable APIs to build their own voice-powered products. He applied to Y Combinator after the official deadline had already passed and was still accepted — a detail that has become part of the company’s origin story.
Under Fox’s leadership, AssemblyAI grew into what is widely regarded as the top-rated API for automatic speech recognition, raising more than $145 million in funding. The company’s models power speech-to-text, speaker detection, summarization, and PII redaction for enterprises including The Wall Street Journal, NBC Universal, and Spotify. Its latest models, including Universal-3 Pro, are built specifically for voice AI applications such as voice agents, medical scribes, and real-time call analysis systems.
Fox has been vocal about his belief that specialization beats multimodal generalists when it comes to reliability — arguing that a model trained specifically on speech-to-text tasks is less likely to produce errors than a general-purpose model handling voice as one of many tasks. He has also argued that managing in-house AI infrastructure is often a costly distraction for companies, positioning AssemblyAI’s specialized infrastructure as a competitive advantage in an increasingly crowded voice AI market.
