Download my resume here!
2023: Reinforcement Learning with Human Feedback Research Project, Toyota Technical Institute at Chicago
This work seeks to explore the fundamental knowledge transfer that takes place when we train
large language models using proximal policy optimization reinforcement methods, specifically in
Reinforcment Learning with Human Feedback. As a part of PPO, a copy of the model before RL training
is made, which we call the "reference model", and used to incentivize the model to maintain the
semantic performance of the original model and prevent reward hacking, a common issue with RL.
The function of this reference model is to be an untrained (and thus not computationally expensive)
copy of the model in which we tether our policy model to. The following quesiton arises almost
naturally: "What happens when we tether our policy model to a more capable reference model?"
This work aims to begin the discussion around this question, and a clear connection to knowledge
distillation research in NLP becomes apparent upon investigation. This project is under advisement by Dr. Kartik Goyal of TTIC
and is being conducted with computational resources supplied by the institute. This work is detailed
on my projects page.
2022-2023: Self-Supervised Learning Research Project, University of Chicago
This work demonstrated the viability of denoising autoencoders as a novel self-supervised learning architecture which surpassed some of the key issues with leading contrastive methods. This work became my M.S. Thesis work and was done under the advisement of and with computation resources provided by Dr. Yali Amit of the University of Chicago Statistics Department. This work is detailed on my projects page.
2022: Data Scientist Intern, The Lubrizol Corporation
At Lubrizol, I focused on a variety of machine learning and analysis projects. These projects were generally centered around sparse chemistry datasets that were extremely wide and short. For this reason, variable selection methods proved to be valuable to most modeling efforts. Most of my work at Lubrizol included introducing variable selection methods and subsequently producing production-grade modeling code and visualizations for presentation to project stakeholders.
2021-2023: M.S. in Statistics, University of Chicago
My degree at University of Chicago marked my transition from Chemical Engineering to Machine Learning and Data Science. During this degree, I studied theoretical statistics, computer science, machine learning, all with a concentration in deep learning and modern statistical methods. I supplemented my coursework with an internship and two research projects to gain experience implementing these types of cutting-edge models.
2019-2021: Process Automation Engineer, The Dow Chemical Company
In Hahnville, LA, one of Dow's largest production sites, I was tasked with creating centralized control code for the maintaining of several key assets. This involved managing control-code repositories, designing and deploying production-grade control code, and working alongside project teams to ensure that all project deliverables could be met in a timely and safe manner. This role provided me with technical skills in software development, as well as project management and communication skills. While immersing myself in automated control, I became fascinated in the sister field of machine learning and decided to turn my attention to learning the fundamentals of data science and machine learning.
2018: Production Engineer Intern, The Dow Chemical Company
At Dow Chemical's emulsions plant in Croydon, PA, I focused on reactor maintenance and altering control strategies to optimize batch performance and eliminate wasteful flushing. I developed an interest in automated control strategies and decided to pursue post-grad work in that field.
2015-2019: B.S. in Chemical Engineering, Villanova University
My coursework and research focused on high-level mathematical modeling within chemical engineering as well as some computer science.
Python (advanced) | R (advanced) | Bash (intermediate) | SQL (intermediate) | Java (beginner) | Django (beginner)
Machine Learning | Deep Learning | Computer Vision | Natural Language Processing | PyTorch | Statistics | Git | Tableau | Power BI