Case Study: Contextual AI accelerates AI model training with WEKA

Contextual AI Builds Production-Ready Enterprise AI Using Google Cloud and the WEKA Data Platform

Contextual AI, a Mountain View company building an enterprise AI platform powered by RAG 2.0, needed to scale AI model training while overcoming long data load times, slow checkpoint writes, and storage limitations in Google Cloud. As its workloads grew across large GPU clusters, its existing Google Filestore-based setup became a bottleneck, hurting developer productivity and delaying model development.

Contextual AI implemented the WEKA Data Platform on Google Cloud, working with Accenture to validate it in a proof of concept before rolling it out across its training environment. With WEKA, the company saw 212% higher aggregate read bandwidth, 497% higher aggregate write bandwidth, 70% lower latency, and 4x faster checkpoint times, while also reducing cloud storage costs by 38% and supporting 100 TB of training data at scale.

Contextual AI

Amanpreet Singh

CTO & co-founder