I’ve done a few projects that I can’t say too much about, at least not any sensitive details. But I can provide a short summary of each.
At a company in the health care sector, I built a new model to predict the volume and type of care provided, 1-2 years into the future. I implemented a nearest neighbours type model in PostgreSQL, and a particular challenge was the calculation speed. The new model provides more accurate predictions than the old one, but maybe more important is that it also works for edge cases that could not be handled by the old model.
I built several models to predict the stock market for a hedge fund (as part of a contest). I used nearest neighbours, logistic regression, and neural networks, among other methods. My earnings are tied to the actual results of the fund, and whether my predictions helped or harmed their profits.
During my PhD at Utrecht University, I built models describing food, water and energy use (and how these interact). The models are part of a much larger model that includes climate, economies and policies, and is used for exploring long-term future scenarios.
This one isn’t actually a secret, but I won’t go into details here. If you really want to know, read the thesis.