Social anthropologist with a PhD in computer science, I am a Bayesian Inference Engineer evaluating risks and opportunities using machine learning and causal reasoning.

In contrast to ad-hoc solutions, the strict application of probability theory (Bayesian approach) guarantees inter-subjective agreements in contexts of uncertainty, the foundation of empirical truths. In addition, it allows to focus attention on the real problem since the specification of causal models is done using intuitive graphical methods that are accessible to everyone. Although it is known to be the correct logic for the empirical sciences, its adoption was historically limited due to its high computational cost. Instead of selecting a single hypothesis, the Bayesian approach evaluates each and every hypotheses according to the empirical and formal evidence (data and causal models). In recent decades, however, these costs have been largely overcome thanks to the development of efficient approximation methods and message-passing algorithms between model variables. It is the correct use of these tools that always guarantees risk reduction and optimal decision making. The state-of-the-art skills estimator for the video game industry, which I have published in Julia, Python, and R), is just one example.