Building the Fastest Real-Time Voice AI Models with State-Space Model Architecture
World-class ML research meets real-time voice AI. Choose Cartesia if you want to work alongside Stanford ML pioneers on cutting-edge SSM architecture — but expect startup intensity.
Cartesia was founded in 2023 by Stanford ML researchers Albert Gu, Chris Re, Karan Goel, Arjun Desai, and Brandon Yang. The company builds real-time text-to-speech models using novel state-space model (SSM) architecture — the same research lineage behind Mamba and S4 models. With just ~91 employees, Cartesia hit $17M in revenue in 2024, demonstrating extraordinary capital efficiency. The company has raised $191M from top-tier investors including Kleiner Perkins, Index Ventures, and Lightspeed. Reviews highlight excellent founders, enormous talent density, and fascinating work — but note the fast pace comes with pressure.
Based on 2 Glassdoor reviews (estimated — very low confidence). Sub-scores not available due to limited review data.
Founded by the creators of S4 and Mamba state-space models. Core ML research drives the product — this is an engineering-first company where researchers ship production models.
~91 people with a flat, many-hats culture. Engineers work across the stack from model architecture to real-time inference optimization. Individual contributors have direct product impact.
Real-time text-to-speech and voice AI. The fastest TTS models in the industry, serving enterprise customers and hitting $17M in revenue within a year of founding.