Subject Matter Expert in multi-fluid flow models and numerical methods. Head of the user’s community of the RoSSBi code. • I co-authored a study explaining chondrule formation in shocks produced by planets, where a database of complex numerical models with a 5D parameter space was constructed and ML clustering was used to extract the most dominant factors. • I increased the speed and accuracy of the gravity solver of the RoSSBi code by applying clustering and multivariable regression into a hierarchical tree algorithm coded in C++. • I improved the degree of insight obtained from the numerical simulations after driving the standardization of data management (storage, cleaning, visualization and analysis) through a Python and Matlab automated process used by the team.

I discovered the instability of dusty vortices with numerical simulations, and their role in planet formation (PlanetS NCCR). • I explained the underlying physics of the instability by exploring a large set of numerical models with 5 independent variables and performing statistics and regression toward a time dependent analytical model. • I was the leader of the project course on Computational Science at UZH, where I mentored Master students in real-world projects solved with Python, using modern algorithms (including ML). • I was the teaching assistant of the lecture on numerical methods for astrophysics at UZH (Master level).

Expert on protoplanetary disk fluid instabilities, and CFD codes. • I reduced the uncertainty and the parameter space of two instabilities, by producing and managing numerical models, performing eigenvalue analysis and modal decomposition. • I doubled the computing speed of the RoSSBi code, that I develop since my PhD, to run bigger and more complex models on computer clusters (distributed systems).