Case Study: AssemblyAI achieves faster, lower-cost speech AI model development with Google Cloud and Vertex AI

AssemblyAI Building complex speech-to-text models faster by upscaling to Google Cloud and Vertex AI

AssemblyAI needed to speed up the training of complex speech-to-text models while reducing the cost and limitations of its on-premises storage and processing environment. Using Looker-backed Google Cloud and Vertex AI infrastructure, the company sought a faster, more scalable way to support its AI research and model development.

Looker helped AssemblyAI move to a cloud-based setup that enabled more simultaneous experiments, automatic hardware upgrades, and stronger audit, logging, and data lineage capabilities. The result was a smooth, cost-saving migration that accelerated innovation: AssemblyAI released its Universal-2 model in less than six months after migration and can now train larger models with billions of parameters more efficiently and at lower cost.

AssemblyAI

Ahmed Etefy

Lead Engineer, Data Infrastructure Team