Solon's interests lie in developing statistical machine learning methods applied to medicine and healthcare. His research focuses on tailored Bayesian inference when different misclassification errors incur different penalties. This is often the case when making decisions around treatments with known side effects where model estimation benefits from incorporating this information. This is also relevant in the diagnostic medical context where for example, missing a disease is usually more severe than a falsely detecting a disease.

Solon is also looking at novel ways to incorporate liquid biopsies into the management of cancer. Liquid biopsies - the analysis of tumours using biomarkers circulating in fluids such as the blood - have the potential to change the way cancer is diagnosed, monitored, and treated.

He is particularly interested (risk) prediction modeling, (medical) decision making and computationally intensive methods, and translational genomics.