At AON, I was asked to investigate the possibilities to use machine learning and other data science techniques to predict insurance fraud. First, I investigated how insurance fraud was currently registered and wrote an advise report on how this registration process can be done more efficiently and in a more data-driven way. Secondly, I extracted claim information from their datawarehouse and built an ensemble of multlple machine learning algorithms to predict and detect insurance fraud. Focussing on the development of machine learning techniques to predict and detect insurance fraud.

During this fixed time fixed scope project at Effectory, I showed how machine learning algorithms could add value to their business and clients. I had the responsability to set up and guide my own project. The final deliverables were multiple machine learning models that could predict employee feedback, interpretable and reusable code files and a report of my most important findings and machine learning methodology.

During this fixed time fixed scope project, I used machine learning techniques to predict daily transaction amounts. In this way, ING can scale up their transaction processing and monitoring systems when large transaction amounts are to be expected.