Dylan Fox

Dylan Fox is founder and CEO of AssemblyAI, which sells speech-to-text and audio-intelligence models through developer APIs. He started the company after years of hitting outdated speech vendors while building AI features, and he has kept the product on a Twilio-style path: research-grade models behind a clean API.

GWU meetups into Cisco speech pain

He studied business administration at George Washington University, drifted toward algorithms through Python meetups and coursework in Washington, D.C., and taught himself deeper machine learning and NLP. At Cisco in San Francisco he worked as a senior software engineer on machine learning for collaboration products, including projects that needed speech recognition. Buying speech as a service, he has said, meant dealing with antiquated sales processes and models that lagged the research literature he was reading.

Founding AssemblyAI as a developer API

In 2017 he founded AssemblyAI to expose current speech research the way Twilio exposed communications. In a Unite.AI interview he recalled the founding spark: "What if you could build a Twilio-style API company using the latest AI research that was just much easier for developers to access state-of-the-art AI models for speech recognition, with a much better developer experience." As a solo founder he applied to Y Combinator after the deadline and still made the Summer 2017 batch. The early company concentrated on accurate transcription APIs, then expanded into audio intelligence — summarization, sentiment, entities, topics, content moderation, and related signals layered on the transcript.

Research in-house, API outside

Under Fox the public product line stayed developer-facing: send audio or streams, get text and structured understanding for telephony, video, meetings, and media products. On Unite.AI he was frank about why teams buy rather than build: "Cost and talent are the biggest challenges... The data is expensive to acquire... compute requirements are enormous to train." Customers, he said, "get access to years of research into state-of-the-art AI models for ASR and NLP, all with a simple API."

On making SOTA speech accessible

His stated mission on Unite.AI is to "make State-of-the-Art AI models accessible to developers and product teams at extremely large scale through a simple API." Language coverage, real-time recognition, PII redaction, and related features grew as customer use cases demanded them. The through-line is developer experience and research currency applied to speech infrastructure other companies put inside their own products.