I'm Miguel Cerne, based in Rio de Janeiro. I work at the intersection of software engineering, data engineering, and quantitative finance — three disciplines that, on their own, don't guarantee anything useful. Put together with discipline, they produce systems people can actually rely on.
Currently a Software Engineering Intern at JGP Crédito, working on the infrastructure behind financial-markets data.
Being wrong is expensive in finance, so the process matters as much as the model.
Every non-obvious call — a threshold, a data cut, a model choice — gets written down with its reasoning. Six months later, nobody has to guess why.
Financial data leaks through time easily. Walk-forward validation and explicit leakage checks are non-negotiable, not an afterthought.
Observer patterns, typed signals, clean boundaries — the goal is that adding a feature never means rewiring the core.