Penn ⊇ {Finance ∩ Data Science ∩ Computer Science ∩ Statistics} | * ∈ {Graduate-Level Course}
One of my favorite courses at Penn, taught by one of the best professors I’ve had. A mix of technical and conceptual. We rebuilt probability models from scratch, worked in Bayesian thinking and updating, and rethought retention and customer profiles. The course is honest about machine learning—what it lacks and what it adds. We looked at models like the Dirichlet that show up in research. Time consuming, but approachable math and worth it. If you like quantitative work and aren’t afraid of math, take it.
This course shows how a computer actually works bits, ISA, compilers, memory, pipelines. We wrote assembly and C, built a C compiler that ties into LC-4 machine code, and moved from gates to higher abstractions. Great class if you want to know what’s hidden when you write higher-level code.
Math heavy and theoretical. Less about coding outputs, more about why algorithms work. The professor emphasizes intuition and proofs: kernels, regularization theory beyond just lambda, SVMs from the ground up. He makes a lot of the theory super approachable though, and it makes it easy to digest. If you're interested in doing more ML courses, I would take this instead of CIS5190: Applied Machine Learning.
A good mix of concept and math. Very research-based. The professor brings in his own work and others’. Clear about the strengths and weaknesses of different models and the assumptions people use in practice. Useful for understanding portfolio construction and attribution.
We covered martingales, Markov chains, proofs, and probability at a deep level. Real analysis ideas showed up in the later part of the course. Difficult but worthwhile. Changed how I think about probability and expectation. Great class if you're willing to sit through it being practically all theoretical.
Light coding in Jupyter, but mostly conceptual with some calculations on exams. This class is what got me into machine learning. Seeing real applications made me want to learn how to build the models myself. Professor was great.
You learn by doing. This class was very fun, and engaging. The professor brings in his own research, and I treated every exercise like it was real. It worked. I got much more confident negotiating after this class, and learned plenty.
A solid look at ETA. Covers parts I wouldn’t have thought about before. The professor is a strong resource—both in the class and for anyone actually interested in starting something. Supportive of student founders and has a crazy track record ($$).