Paris, Île-de-France, France
Software Engineer @ Datadog, MScEng @ CentraleSupélec/University of São Paulo. Currently building benchmarking systems at Datadog to make our APM suite faster. Before my industry experience, I was pursuing a specialization on mathematics and data science within CentraleSupélec, where I worked alongside research groups and companies on multiple machine learning projects: - Denoising of X-ray images with General Electric Healthcare/Laboratoire Signaux et Systèmes. - Detection and reconstruction of roads from radar satellite images with researchers from SONDRA/European Space Agency. - Speech separation for automatic tracking speakers with Orange. - Automatic e-commerce product title generation with BearingPoint. My passion for math, statistics, machine learning and computer science was fostered at the University of São Paulo, where I authored an undergraduate research project on deep learning with complex numbers and worked on a SWE internship at BTG Pactual. Notably, I ranked 2nd out of 249 electrical engineering students, 1st out out of 56 in my specialization and was awarded the Eiffel Excellence Scholarship to cover all my expenses while studying in France. Beyond technology and engineering, I have passions in climbing, lifting, calisthenics, fixed gear cycling, TTRPGs, making music, writing. Happy to chat about any of these if you want to break the ice :)
Built a data intake system to ingest customer data for the company's machine learning models.
Enhanced Datadog's benchmarking system to accelerate benchmark development, streamline result visualization, and inform decisions on performance.
- Developed custom complex-valued neural networks in TensorFlow, improving classification accuracy by 10% on complex-valued datasets - a feature not supported at the time of the research in TensorFlow releases. - Published research findings at the University of São Paulo's International Symposium of Undergraduate Research.
- Developed a data merging system that resolved critical credit document issues for over 300,000 customers. - Automated key operational processes, resulting in a time savings of over 100 work hours monthly for a team of 5 people.