Case Study: Contextual AI automates GPU-based CI and accelerates testing with Modal

How Contextual AI automated CI with Modal GPUs

Contextual AI, an end-to-end platform for building RAG 2.0 enterprise AI applications, faced a bottleneck in its development process. Their continuous integration (CI) testing required GPUs to run tests on LLMs, but manually running these tests on in-house resources was inconvenient, time-consuming, and risked broken code. Their existing cloud options were either too slow, unreliable, or could not provide the necessary multi-GPU configurations per job, creating a major hurdle for their growing engineering team.

By implementing a solution with Modal, Contextual AI fully automated its CI workflow. A GitHub Action triggers a Modal Function with multiple GPUs to run their test suites. This provided parallelized testing, rapid startup times due to image caching, and on-demand billing that kept costs low. The results for Contextual AI were maximized developer iteration speed, a higher quality bar, and the elimination of manual testing, all enabled by Modal's infrastructure and excellent technical support.

Contextual AI

Stas Bekman

ML Engineer