Matthew Zackey

Background

About Matthew Zackey

I recently completed my master's in Mathematical Statistics at Stellenbosch University with a Cum Laude. My master’s degree was two years, with modules in the first year and a thesis in the second. My modules included Statistical Learning Theory, Extreme Value Theory and Resampling Techniques. My thesis was on boosting, where I covered AdaBoost, Gradient Boosting, XGBoost, LightGBM and CatBoost. I also presented part of my thesis at the South African Statistical Association (SASA) conference at the end of November last year.I am interested in machine learning and AI, specifically deep learning, reinforcement learning, computer vision, diffusion models, LLMs and multimodal models. What captivates me the most about machine learning is looking under the hood of a model and understanding its algorithmic and mathematical details. Additionally, I find extreme satisfaction in taking theoretical concepts and translating them into practical implementations through coding. Currently, I am working on training RL agents in OpenAI gymnasium. My completed projects include Deep Q-learning with Experience Replay and applying Deep Deterministic Policy Gradients. These projects are available on my GitHub. Additionally, I have been working through LeetCode to improve my knowledge of Data Structures and Algorithms. I have currently completed 60+ questions. This year, I aim to work in machine learning at a company or pursue a PhD in machine learning.Links- GitHub: https://github.com/MattZackey - LeetCode: https://leetcode.com/Mattzackey/Programming Languages- Python- R programmingLibraries and Frameworks- PyTorch- NumPy- Pandas- Scikit-Learn- XGBoost- LightGBM - CatBoost