You know XGBoost, but do you know NGBoost? I'd passed over this one, mentioned to me by someone wanting confidence intervals in their classification models. This could be an interesting paper to add to the ML curriculum.