I'm an Applied Scientist in Artificial Intelligence at Amazon.
I design and deploy AI systems that solve real-world problems at scale,
and I bring my scientific approach to Machine Learning research.
I earned my PhD in Physics from Carnegie Mellon University.
I worked in observational Cosmology as an Analysis Team Lead
in the Dark Energy Survey.
I love dancing, lifting, learning, and traveling.
At Amazon, I build generative AI applications for FinTech use cases. My work spans novel research in LLM-based text compression, intelligent document analysis using retrieval augmented generation, and machine learning systems processing millions of records. I bring a research-driven mindset to every project, from algorithm design through production deployment. I also serve as a peer reviewer for Amazon's internal science conferences.
I earned my PhD from Carnegie Mellon, where I developed AI-driven methods for analyzing massive astronomical datasets. My thesis, "Dark Energy Science from 100 Million Galaxies," focused on machine learning pipelines for galaxy distance estimation and a novel empirical approach to model selection that has since been adopted by other leading research collaborations.
I spent five years as a member of the Dark Energy Survey, one of the most ambitious cosmological surveys ever undertaken. I served as Analysis Team Lead, directing teams of scientists studying the accelerating expansion of the universe. I developed machine learning methods that significantly improved galaxy distance estimation, and I earned the DES Builder distinction for sustained infrastructure contributions.
At the Institute of Theoretical Physics in São Paulo, I co-led a large-scale Bayesian analysis that achieved a two-fold improvement in cosmological parameter estimation. This work was published in Physical Review D. Unesp is where I sharpened my skills in advanced statistical methods and high-performance computing, bridging theoretical physics with data-intensive research.
This is where it all started. I earned my Bachelor's in Physics from UFSJ in Brazil, a program with a strong emphasis on computational and theoretical methods. From my earliest semesters, I was trained in numerical algorithms, scientific computing, and mathematical modeling. This foundation in computational physics set the trajectory for everything that followed, from cosmological simulations to the AI systems I build today.
When I'm not working on research, I enjoy teaching,
building communities, and advocating for diversity in STEM.
I've taught physics at multiple levels, as a University Lecturer at UFSJ in Brazil and as a Teaching Assistant at Carnegie Mellon. I also co-organized the McWilliams Software Development Series, running monthly tutorials for the research community.
I'm an active participant in Women in Data Science (WiDS) and the Women in ML Symposium. I completed the Data Science For All / Women program and am passionate about building inclusive communities in science and technology.
I served as President of the Tartan Salsa Latin Dance Club at CMU, managing a 10-officer board, securing funding, and launching cultural initiatives that strengthened community engagement on and off campus.
I serve as a peer reviewer for Amazon's internal science conferences, evaluating research submissions for scientific rigor, novelty, and practical impact across multiple conference cycles.