Case Study: Alis Exchange achieves lower AI model costs and ultra-low latency with Google Cloud Platform

Rezco and Alis Exchange Building ML models at scale, with clearly defined cloud architecture

Alis Exchange, a Luxembourg-based software company working with asset manager Rezco, needed to build an end-to-end AI system that could analyze large volumes of financial market data and support portfolio research at scale. As its Alis Alpha platform grew more complex, the team turned to Google Cloud Platform and its cloud engineering approach to bring order, scalability, and consistency to the system.

Using Google Cloud Platform services and SRE principles, Alis Exchange restructured its platform into Alis Build and rebuilt Alis Alpha with tools such as Cloud Run, Cloud Dataflow, Cloud Bigtable, and TensorFlow. The result was a far more efficient system: model-training costs dropped from $800 to just over $100, latency improved from 15 milliseconds to 2 milliseconds, and the team gained 10x faster and cheaper API calls, enabling faster experimentation, smoother team onboarding, and more powerful market forecasts.

Alis Exchange

Jan Krynauw

Chief Executive Officer