Felipe C. Pacca

About

Felipe C. Pacca works in AI engineering, ML & data science and AI automation & agents, based in Rio de Janeiro - Brazil. B.Sc. Applied Math · AI Engineer · Founder of Solaris — a company staffed by AI agents: support, outbound, marketing, onboarding; 7 agents, 1 human. Public projects on GitHub include Machine Learning From Scratch (Pure NumPy implementations of core ML algorithms: Linear/Logistic Regression, Decision Trees, K-Means,…), Titanium Alpha (Agentic multi-strategy hedge fund: PatchTST forecasts, 4-agent LangGraph debate, CPCV-OOS + DSR validation,…) and Ibov Data From B3 (Project aimed at gathering, standardizing, and exploring historical Ibovespa data obtained from B3,…), mostly written in Jupyter and Python.

Projects 6

• 

  Machine Learning From Scratch

  Pure NumPy implementations of core ML algorithms: Linear/Logistic Regression, Decision Trees, K-Means, Neural Networks, Recommender Systems, and RL. Focused on mathematical derivation and vectorized optimization.

  github.com
• github.com

  Titanium Alpha

  Agentic multi-strategy hedge fund: PatchTST forecasts, 4-agent LangGraph debate, CPCV-OOS + DSR validation, HRP with Ledoit-Wolf shrinkage. 10-year OOS Sharpe=0.766.

• 

  Ibov Data From B3

  Project aimed at gathering, standardizing, and exploring historical Ibovespa data obtained from B3, preparing a consistent basis for quantitative analyses, automation, and publication of results.

  github.com
• 

  Stanford Machine Learning Course

  Python notebooks from Stanford’s Machine Learning specialization, covering supervised learning, unsupervised techniques and advanced algorithms.

  github.com
• github.com

  Ibm Ai Engineering Course

  Python notebooks from IBM’s AI Engineering program covering ML, deep learning, neural networks, computer vision and generative AI.

• 

  Breast Cancer Prediction

  This project demonstrates how to build, train, and save a deep learning model for breast cancer prediction using TensorFlow and Keras.

  github.com