I work seamlessly across descriptive and inferential analysis, statistical modeling, Business Intelligence and data product development. My solid background in software engineering informs my CI/CD practices and enables me to contribute smoothly to Agile teams.

My expertise encompasses time-series analysis, classical machine learning, deep learning and large language models (LLMs), as well as complex interactive visualizations built with React and d3.js. For more than 3 years I’ve worked with Microsoft Azure cloud services (including ADF, Databricks, Azure ML and Power BI) while specializing in SQL query optimization and advanced Python programming using frameworks such as Flask, FastAPI, scikit‑learn, PyTorch and Streamlit.

Since 2015 I have been immersed in the open‑source and startup ecosystems, collaborating with high‑performing, cross‑functional teams. Over the past 3 years I have contributed to data teams in complex industrial settings, focusing on anomaly‑detection systems, recommendation engines and computer‑vision projects: boosting operational efficiency, advancing Industry 4.0 initiatives and delivering measurable profits.

Beyond my professional roles, I spearheaded ARBOSOCIAL: an open-source arbovirus prediction system data-lab-org/ARBOSOCIAL. By integrating social, climate and health data, building ensemble ML models in PySpark and scikit-learn, and surfacing results in interactive notebooks, we achieved over 85% accuracy and supported early-warning efforts in pilot regions.

🔗 Resume: kelly.decastro.com.br

🔗 Stack Overflow: users/18042741/kelly-de-castro

🔗 Current focus: data-lab.org

Let’s connect!