Researcher and data scientist with a background in Social Sciences,
an ongoing PhD in
Political Science (DCP-USP),
and an MBA in Data Science & Analytics (ESALQ/USP).
Currently finishing my PhD and open to data science roles in industry.
My research investigates how Brazilian politicians use social media
to communicate positions on public health — combining large-scale
data collection via APIs, Deep Learning models (BERT), and political
discourse analysis. I bridge tools that rarely go together: NLP with
political science, academic rigor with production-ready code.
My core project builds and analyzes a corpus of Brazilian legislators'
tweets on COVID-19 vaccines (2020–2023), annotated for stance and
sentiment — published in the ACL Anthology (NLP4DH 2025) and presented
at STIL 2023.
I also develop teaching materials when existing ones fall short: I built
an interactive textbook
on Python, NLP, and Machine Learning for social science researchers
without a technical background, with executable notebooks via Google
Colab.
Featured Projects
Stance Detection with BERT
Fine-tuning BERT models to automatically classify political stance
on COVID-19 vaccines in tweets by Brazilian legislators.
Published at STIL 2023.
Dataset of 9,045 posts annotated for stance and sentiment from
Brazilian political elites on COVID-19 vaccines. Publicly available
at ACL Anthology (NLP4DH 2025).
Interactive textbook on Python, Machine Learning, and NLP built
for social science researchers without a technical background.
Executable notebooks via Google Colab.
I started in academic research through empirical political science —
but each project pushed me deeper into data. At the Rede de Pesquisa
Solidária (2020–2025), I learned to work with large-scale government
data and produce analyses that inform real decisions, in collaboration
with USP and FGV researchers. At Instituto Butantan (2021–2022), I built
public policy databases from official government sources. At Fundação
José Luiz Egydio Setúbal (2023–2025), I went deep into NLP: collecting
via API, training BERT models, and publishing results at international
conferences.
These experiences shaped a profile that values crossing boundaries:
between social sciences and machine learning, between reproducible
research and practical deliverables, between methodological rigor and
real-world impact.
Contact
I'm open to conversations about data science roles, applied research
collaborations, and projects involving NLP, text analysis, and
unstructured data.
