Paulo Loyola

Co‑developed an internal retrieval‑augmented LLM assistant (Llama‑3 + Vertex AI + LangChain) to accelerate access to company documents and led design of a forecasting system that supported real‑time campaign decision‑making.

Enhanced auto fraud detection with explainable GLMs and integrated models into real‑time claims pipelines; developed NLP classifiers and customer propensity models to optimize cross‑sell campaigns.

Designed ARIMAX/SARIMAX forecasts for supply‑demand planning and rapidly prototyped feature engineering pipelines to meet client constraints in retail and logistics.

Researched ML metrics to evaluate stock performance and detect trading anomalies using unsupervised learning and contributed to fraud detection prototypes on trade data.

Pursuing a Master of Science in Computer Science with coursework and research in machine learning and scalable ML systems.

Completed the Deep Learning Nanodegree focused on advanced deep learning techniques and practical projects.

Completed a course on Generative AI with Large Language Models covering LLM architectures and applications.

Completed the Deep Learning Specialization covering neural networks, CNNs, RNNs, and practical deep learning workflows.

Completed the Machine Learning course covering supervised and unsupervised learning algorithms and model evaluation.

Completed a Bachelor of Applied and Computational Mathematics with focus on statistical modeling and data science applications.

Completed the Statistical Learning course covering statistical methods for modeling and inference.

Completed an online course in SQL focused on database querying and data manipulation techniques.