Karan Goel

Also known as krandiash

Karan Goel is a computer scientist and entrepreneur whose research on state space models grew into a voice AI company. He completed a machine learning master's degree at Carnegie Mellon University in 2018 and a computer science PhD at Stanford, advised by Christopher Ré. With Albert Gu and Ré, he coauthored the S4 sequence model, and ICLR gave the paper an Honorable Mention in 2022. In 2023, he co-founded Cartesia with Gu, Arjun Desai and Brandon Yang, and the company raised a $64 million Series A in March 2025 and a $100 million round in October 2025. He is CEO and founder of Cartesia, and in June 2026 he pitched the company as both the speaking and listening layer for real-time AI agents.

Now at: Cartesia
Badges: Founder
Bridge
Links: Google Scholar, X, LinkedIn and GitHub

Timeline

2018

Carnegie Mellon and a Siebel Scholarship

He studies electrical engineering at the Indian Institute of Technology Delhi and completes a master's degree in machine learning at Carnegie Mellon University's Department of Machine Learning in 2018. The Siebel Scholars Class of 2018 names him among its scholars from Carnegie Mellon's School of Computer Science.
2021

Robustness Gym for NLP evaluation

At Stanford, where he earns a computer science PhD, he is advised by Christopher Ré and affiliated with the Stanford AI Lab. The paper Robustness Gym, which he writes as first author and posts to arXiv in January 2021, presents an NLP evaluation toolkit that unifies subpopulations, transformations, evaluation sets and adversarial attacks.
2022

An ICLR honorable mention for S4

ICLR 2022 gives an Honorable Mention to "Efficiently Modeling Long Sequences with Structured State Spaces," which he coauthors with Albert Gu and Christopher Ré. The paper, posted to arXiv in October 2021, introduces the Structured State Space sequence model, or S4, which reaches state-of-the-art results on every Long Range Arena task.
2023

Co-founding Cartesia

After part-time jobs at Snorkel AI and then Salesforce, he joins Albert Gu, Arjun Desai and Brandon Yang in 2023 to launch Cartesia. The startup emerges from Stanford's AI Lab to commercialize their state space model research.
May 2024

Sonic, a low-latency voice model

In May 2024, he announces Sonic, Cartesia's low-latency voice model that generates lifelike speech. His launch post says the company's models build on state space models such as S4 and Mamba, which its team developed in academia with Tri Dao.
December 2024–March 2025

Funding from Index and Kleiner Perkins

A $22 million round led by Index Ventures, announced in December 2024, brings Cartesia's total raised to $27 million. In March 2025, he announces a $64 million Series A led by Kleiner Perkins together with the Sonic 2.0 voice model, and Cartesia has raised $91 million in total.
October 2025

A $100 million round and Sonic-3

In October 2025, Cartesia closes a $100 million round, with Kleiner Perkins, Index Ventures, Lightspeed and NVIDIA named as investors. He launches Sonic-3, a real-time text-to-speech model built on state space models that supports 42 languages.
As of October 2026

Leading Cartesia as CEO

He is CEO and founder of Cartesia, a startup building multimodal foundation models for real-time applications. In June 2026, he pitches Cartesia as both the speaking and listening layer for real-time AI agents, with its Sonic-3.5 text-to-speech and Ink-2 speech-to-text models.