Angikar Ghosal

Class of 2024
Majors in Mathematics and Computer Science, Artificial Intelligence/Machine Learning concentration; Minor in Economics

I currently perform research in generative AI under Professor Bhuwan Dhingra and theoretical machine learning under Professor Rong Ge. I have been a teaching assistant/grader for many classes across computer science, mathematics, economics and statistics: twenty four and counting (six in computer science). I am Head of Data Science and Recruitment at DAML, Duke's largest student tech organization, and write for Juris, the undergraduate law magazine. I worked at Barclays, the British bank, in the Sales & Trading division in Spring 2023 and rotated on Rates Options, Non Agency Securitized Products and Equity Derivatives. My hobbies include reading (e.g., Dante, T. Mann, Mishima, J. Baldwin, J.L. Borges, more recently Krasznahorkai), classical music (e.g., Mozart, Puccini, Vivaldi, Wagner, more recently Glass, Gorecki), and films (e.g., Kieslowski, Ozu, F. Coppola, more recently Audiard).

Favorite Class: Favorite class overall: Physics 627 (Quantum Computing), as it had exceptionally well-designed problem sets, and the pedagogy involved charting your own path through the subject, incorporated approaches from computer science, physics, mathematics, statistics, chemistry, and mostly because quantum mechanics is the most beautiful theory ever invented by mankind. Favorite computer science class: Computer Science 689 (formerly 590) (Advanced Deep Learning), as the homeworks involved implementing actual research papers from the last 5 years, and the professor aesthetically linked together deep learning algorithms with inspirations from statistical thermodynamics, solid state chemistry, graph theory. Favorite math class: Math 501 (Abstract Algebra I), as it taught me to look at the world in a new way (through symmetries), introduced me to the rigor of abstract reasoning (which still had lots of real-life applications) but most importantly, kept me on my toes by making me revisit reasonings of prior claims and lemmas again and again.