People trained in the same field tend to agree with each other, including about what they overlook. Sofia had seen this often in research, and in 2026 she measured it in an independent study where AI reviewers with different backgrounds caught more problems than identical ones, and still shared one blind spot. A second opinion is most useful when it comes from a different training, and that is the idea CosmoVentures is built on.
Sofia studied astrophysics at ETH Zurich, Caltech and Université Paris Cité, observing the faint gas between galaxies with the large telescopes in Chile and Hawaii and with Hubble. Part of that work was published in Nature. Alongside research she has kept working in music, games, workshops and writing, and over the years these have become part of how she approaches any new problem, as much as the science has.
This is the thread that connects the work on this page. A technical claim gets checked against the evidence. A dataset or an AI tool gets tested on real cases. A plan gets read from outside the field that wrote it. An explanation gets tried on the people it was written for. In each case the method is the one research taught, and the subject is whatever the client brings.