Hello! My name is Neeltje Kackar. I am a Senior (4th-year) undergraduate student at William & Mary studying Data Science and Physics. I am very interested in computational physics, machine learning, and scientific research.
Currently, I am working as a Research Assistant in the William & Mary Physics Department under the supervision of Prof. Saskia Mordijck. My research focuses on using machine learning techniques to predict $\tau_e$ scaling in tokamak reactors. I've worked with a lot of different subfields during this work, covering topics such as high-performance computing, attentive neural networks, variational autoencoders, reinforcement learning, uncertainty quantification, interpretable ML, and more.
In addition to my research, I am also a Teaching Assistant for the Data Science Department at William & Mary under Prof. Daniel Vasiliu, where I help with DATA 301: Applied Machine Learning. I write class materials, assist with writing the homeworks and exams, hold office hours twice a week to help students, and answer students' questions online. In this capacity, I've worked with structured embedding models for Python code, as well as finetuning LLMs. I'm also working on a senior honor thesis on predicting airplane delays across whole fleets using Graph Neural Networks.
In my free time, I enjoy working on personal coding projects, and watching YouTube lectures/documentaries. I also founded 1991 Linux Club at William and Mary, and have been its president for 2 years now. I mainly use Ubuntu and Debian Sid. I always enjoy learning new things and strive to be as well-rounded as possible, while still going into depth on important topics!
I am graduating in May 2026 and am currently applying to graduate schools, with my primary foci on Optimizing Machine Learning Algorithms, Improving LLM Performance, and AI for Computational Physics. If you have any questions about me or my work, feel free to reach out! I check my email multiple times per day and respond quickly: grad@neelt.je.