For a global hedge fund, we developed a systematic trading strategy using SOTA techniques in Machine Learning coupled with financial markets microstructure knowledge. The results were pretty decent in the American equities market. 

For this client, we re-constructed internet users' journeys and applied data analysis to decrypt their online behaviour (browsing, searching and streaming activities). Our predictive algorithms enable the client to come up with tailored product recommandations.

Probability of default (PD) and loss given default (LGD) are the two central elements of credit modeling. We succeeded in beating the client's exisiting mathematical models by using the proper machine learning algorithms and data features. The result is an increase of speed and accuracy in pricing insurance contracts.

Recommandation engines are powerful tools for players in the Digital Advertising industry (AdTech). Using SOTA Transfer Learning algorithms helped our clients sharpen their products recommandations by extracting optimal textual and visual features from a (very) large products catalogue.