Case Study: Lightricks achieves scalable recommendation engine deployment with Qwak

Lightricks launches recommendation models in remarkably short time with Qwak

Lightricks, a developer of video and image editing mobile apps including Facetune, faced challenges in scaling its machine learning operations. The company needed to build a complex recommendation engine based on tabular data, which required daily training and significant engineering support to build new infrastructure layers. They sought a flexible MLOps platform to avoid vendor lock-in and to centrally manage training and deployment at scale without the immense effort of building it in-house. Other vendors could not meet their requirements for ease of use, onboarding speed, and feature store support.

Qwak implemented its end-to-end MLOps platform to address these challenges. The solution provided Lightricks with capabilities for model training on fresh data, simple deployment with canary releases, a feature store for autonomous data science work, and model monitoring with real-time alerts. Using the Qwak platform, Lightricks was able to deploy its complex recommendations solution within a remarkably short timeframe. The partnership with Qwak enhanced the company's capacity to handle the complex demands of modern AI/ML pipelines.

Lightricks