Skip to content
View JF-Andrade's full-sized avatar

Highlights

  • Pro

Block or report JF-Andrade

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
JF-Andrade/README.md

M.Sc. Economics | Machine Learning & Causal Inference

LinkedInPortfolioEmail

Data Scientist and Master in Economics with over four years of experience leading strategic machine learning and market intelligence projects. I specialize in bridging the gap between advanced Econometrics and scalable Machine Learning to unlock business value through uncertainty quantification and causal discovery.

Technical Expertise

Statistical Modeling
Causal Inference Bayesian Econometrics

Machine Learning
TensorFlow PyTorch Scikit-learn

MLOps & Infrastructure
MLflow Docker AWS

Methods
DID RDD Synthetic Control A/B Testing Time Series Hierarchical Modeling

Tools
Python SQL Pandas NumPy CatBoost Optuna PyMC Git


Strategic Projects

Marketing Science: Bayesian MMM

Impact: Uncertainty Quantification (94% HDI) for multi-region budget optimization.

Decision-support system estimating channel ROI across 18 territories to maximize advertising efficiency. Delivered as an interactive Budget Optimization Engine (Streamlit) with executive reporting and a reproducible MLflow pipeline. Uses Geometric Adstock and Logistic Saturation to isolate true media effects.

  • Stack: PyMC-Marketing ArviZ Streamlit

View Code

Personalization: Recommender Systems

Impact: +389% Lift in purchase conversion vs. naive baselines.

Large-scale retrieval & ranking funnel serving 1.37M customers. Delivered as a production-grade Two-Stage pipeline with MLOps tracking and visual embeddings (ResNet50), built to handle seasonal catalog rotations and cold-start products.

  • Stack: TensorFlow Recommenders CatBoost Optuna

View Code


Academic & Publications

  • Master of Science in Economics — UNICAMP (2023)
  • Bachelor of Economics — UFRJ (2017)

Publication: Andrade, J. F., et al. (2026). Investment share and economic growth in Latin America. Review of Keynesian Economics.


Pinned Loading

  1. JF-Andrade.github.io JF-Andrade.github.io Public

    Personal portfolio and Data Science showcase featuring bilingual content (EN/PT) and production-grade ML case studies. Built using HTML, Tailwind CSS, and Vanilla JS.

    HTML

  2. two-tower_recommendation two-tower_recommendation Public

    End-to-end Two-Stage Recommendation Architecture (Two-Tower Retrieval + CatBoost Ranking) for H&M Personalized Fashion. Engineered as a Python package..

    Python