Some examples of our projects:

 CDN infrastructure improvements 

Deployed a Deep Learning model that automated the detection of CDN infrastructure bottlenecks. Improvements helped Telecom customer increase revenue in some regions (Brazil, Taiwan, Florida) and become the best-performing CDN (based on internal and 3rd party metrics).

Technology: Spark, ClickHouse, python, keras, tensorflow, CatBoost, airflow

 Sensors failure detection

Automated anomaly detection workflow with a deep learning approach on image data (in the spectral domain) that reduced project costs and decreased the time needed to fix equipment issues.

Technology: python, tensorflow, keras, ResNet50V2, docker, AWS

 Automated EarthQuake QC process

Delivered Machine Learning classification model that distinguishes fake and real seismic activities. Projects are completed faster which leads to a significant cost reduction. Elimination of manual touchpoints resulted in quality improvement as well as more stable outcomes.

Technology: matlab, python, scikit-learn, AWS

Case Studies