About
Hi! I am Rafael Zimmer, an undergraduate CS Student at USP.
As of now, I am mostly focused on my studies, currently developing a suite for quantitative analysis tasks, called dxlib, as well as deploying strategies to production (go take a look at the Forge project!
Lately, I've developed quite the interest for Reinforcement Learning, specifically for financial agent optimization. Have you ever heard of Jax? Or Flax? Seem very useful!
News
- Aug, 2023: Published the paper Hybrid Models for Facial Emotion Recognition in Children in the CLIHC 2023 Congress 🎉
- Jul, 2023: Initial 1.0 release of the dxlib library! 😎
- Jan, 2023: Joined my first internship in finance, as a Quantitative Analyst @ Clave.
Research
I've mostly developed code and research for Computer Vision tasks, focusing on classification and segmentation tasks. Some very interesting projects include segmentating rivers and building clusters for drone navigation. For some time, I've also been part of the Hugging Face Ambassador's team, creating tutorials and content on how to use the library, as well as the features of easily importing and running ML models. If you'd like to know more, access the Education Toolkit for HF.
More recently, my interest has turned to financial markets, specifically quantitative and algorithmical trading. My current focus is on using RL algorithms for market making strategies.
Publications
-
Optimizing a Market-Making strategy using Reinforcement Learning
Authors: Zimmer, R.
Published: arXiv:cs.LG, 2024
Unfinished -
Machine Learning Techniques for Improving Multiclass Anomaly Detection
Published: IEEE I2MTC 2024
Code -
