Amin Rahimian, Teaching

My teaching ranges from required undergraduate courses to master's core courses, Ph.D. qualifiers, and advanced doctoral electives, spanning statistics, the design of experiments, stochastic processes, randomized algorithms, and AI. Some of my teaching philosophy took shape while working with Santosh S. Venkatesh at the University of Pennsylvania, across his undergraduate and graduate probability courses and his open probability MOOC. From that experience I came to prioritize clarity of exposition and to treat problem-solving as the organizing principle. In my own courses, I build on this by introducing each method through the concrete question it answers and the real-world choices and designs it informs, and by working to move students from executing analyses to asking better questions of their own.

AI for Social Good

A hands-on introduction to machine learning and modern AI through project-based work with real-world data — in domains such as public health, consumer credit, and online platforms — paired with readings on fairness, privacy, accountability, and data ethics. Redesigned around mastery-based learning and LLM-integrated tutorials that build on a guided technical core with optional deeper extensions, and combine coding with model auditing, output verification, and reproducible GitHub workflows, the course asks students to weigh both the power and the responsibility of using AI in their own analyses. It supports the Engineering Data Analytics Certificate and Pitt Business's Certificate in Foundations of AI for Business. Formerly Data for Social Good. Public materials: github.com/​aminrahimian/​ai-for-social-good.

Foundations of Statistics

A first-year Ph.D. course in mathematical statistics, developing statistical inference across its frequentist, Fisherian, and Bayesian traditions. Core topics include maximum-likelihood and minimum-variance unbiased estimation, sufficiency, Fisher information and the Cramér–Rao bound, hypothesis testing and the Neyman–Pearson framework, likelihood-ratio tests, interval estimation, Bayesian inference (with priors, loss, and risk), decision theory for evaluating statistical procedures, and the linear model through ANOVA, experimental design, and regression — with attention to how these classical foundations meet modern machine-learning and data-driven paradigms. It supports the industrial engineering statistics Ph.D. qualifying exam.

Statistics and Data Analysis

A master's course in statistical modeling and data analysis, taught in R. It builds from probability and sampling foundations through estimation, confidence intervals, and hypothesis testing to analysis of variance and factorial experimental design, and develops regression as its central tool for modeling and prediction. It also draws on mathematical statistics to treat likelihood-based and Bayesian estimation, giving students both frequentist and Bayesian ways to reason about data — with a consistent emphasis on applying these methods to real datasets and interpreting what the results support.

Probabilistic Analysis of Algorithms and Randomized Algorithms

An advanced Ph.D. elective on the use of randomness in algorithm design and the probabilistic analysis of algorithm performance, as the principal points of contact between probability and computation. The first part develops the core probabilistic tools: concentration bounds, the union bound, the probabilistic method, coupling, martingales, and random graphs. It applies them to the analysis of randomized algorithms, Monte Carlo estimation, sampling, counting and Markov chain Monte Carlo, differential privacy, sample complexity, and statistical and computational learning theory. The second part covers seminal papers on decision-making under uncertainty: optimal stopping, multi-armed bandits, and learning on networks. The course pairs technical mastery with critical reading of research, culminating in student-led paper presentations. Public materials: github.com/​aminrahimian/​probabilistic-​analysis-​of-​algorithms-​and-​randomized-​algorithms.

Design of Experiments and Quality Assurance

A required undergraduate course on the technical foundations of designed experiments and statistical quality control. The course sets each topic in its practical and historical context (from Fisher's founding of experimental design to Deming's quality revolution in industry) and develops the statistical methods that the applications require. The material spans probability and statistical inference — sampling, estimation, hypothesis testing, and statistical power — then the design of experiments, including analysis of variance and factorial and fractional-factorial designs, and finally industrial quality control through acceptance sampling and control charts. R is used throughout, with no prior experience assumed. Public materials: github.com/​aminrahimian/​design-​of-​experiments-​and-​quality-​assurance.

Stochastic Processes

A first-year Ph.D. course covering probabilistic modeling from its foundations through the main classes of stochastic processes. Topics include conditioning and expectation, the modes of convergence of random variables and the limit theorems (the laws of large numbers and the central limit theorem), discrete- and continuous-time Markov chains, the Poisson process and the exponential distribution, renewal theory, and queueing theory, with applications across operations, reliability, inventory, and communication systems. Only elementary probability and calculus are assumed, and the course supports the industrial engineering stochastic-processes Ph.D. qualifying exam.