Projects

Selected projects

Data products, research software, and applied ML work.

A focused selection of projects that translate academic depth into industry-relevant delivery: model design, experimentation, production thinking, and stakeholder communication.

E-commerce recommendations

Recommendation Systems at Mercadolibre

Production-scale machine learning for ranking and recommendation in a major e-commerce marketplace.

  • Role: Senior Data Science Machine Learning Engineer.
  • Work: ranking models, heterogeneous feature engineering, fashion attributes, logistics signals, collaborative interactions, embeddings, MLflow workflows, A/B testing, rollout, and monitoring.
  • Industry value: model quality, reproducible experimentation, production ownership, and cross-functional delivery.

Applied data science leadership

Data Science Lead at Ithreex Global

Led and delivered data science projects across agriculture, finance, retail, and tourism.

  • Role: Senior Data Scientist and Data Science Lead.
  • Work: computer vision for livestock weight estimation, customer segmentation, churn modeling, forecasting, RAG-based support systems, and internal Python libraries for model deployment.
  • Industry value: translated ambiguous business problems into deployable models and reusable engineering assets.

Scientific software

Open-Source Research Software

Scientific software and data-intensive pipelines developed through astronomy research collaborations.

  • Projects: HEARSAY, AEGIS, PINNACLE, CORRAL, PROPERIMAGE, GriSPy, and related public repositories.
  • Work: statistical simulation, publication analytics, astronomical image processing, pipeline design, reproducibility, and package documentation.
  • Industry value: production-minded software practices in research environments with complex data and long-term reproducibility needs.

Public-interest analytics

ARCOVID19 Decision-Support Work

Multidisciplinary data work during the SARS-COVID-19 pandemic.

  • Work: decision-support tooling, data analysis, public communication, and collaboration with scientific and institutional teams.
  • Industry value: fast analytical delivery under uncertainty, explainable results, and communication with non-specialist stakeholders.

Astronomical event pipelines

TOROS and Transient Detection

Research and software contributions to optical follow-up of gravitational-wave events and transient astronomical sources.

  • Work: image analysis, machine learning for real/bogus detection, data pipelines, and collaboration with international teams.
  • Industry value: automation over noisy data streams, model evaluation, operational reliability, and distributed scientific collaboration.