I'm a software engineer interested in machine learning and robotics. I graduated from the University of Maryland (May 2025) in Computer Science with a ML specialization and minors in mathematics and robotics. I currently work at Comcast on data infrastructure as a software engineer.
Projects
I'm currently working on building a transformer end-to-end from scratch to learn more about the rigorous process of training large neural networks, scaling, and getting to learn about the fundamentals in a practical way. 1This originally started with the Karpathy lectures, but turns out the training infrastructure, data pipeline efficiency, and ablation tests were the hardest (but most interesting) parts. Blog post on that soon. This includes a custom BPE tokenizer, data pipeline, and training infrastructure with only PyTorch, implementing several modern architecture improvements (RoPE, FlashAttention, GQA, etc) ontop of the base decoder model. You can find the repo here (opens in new tab).
At the University of Maryland, I built the core ROS2 infrastructure for an autonomous rover project and designed the website (opens in new tab) and branding for Robotics at Maryland, the university's largest student-run robotics organization.
Contact
You can contact me at [email protected].