-
Lally 316
- gittea@rpi.edu
-
518-276-6476
- 0000-0003-3482-0157
About
Before joining Rensselaer as an Assistant Professor of Computer Science in 2017, Dr. Gittens was a research scientist at eBay Research Labs and a postdoctoral scholar at the International Institute of Computer Science and the Department of Statistics, UC Berkeley. His interests lie in applying randomness to computational linear algebra, machine learning, and statistics, and are motivated by the goal of efficiently extracting information from massive data sets while maintaining guarantees on rigor and accuracy.
PhD in Applied and Computational Mathematics, Caltech, 2013
Industry Postdoc, eBay Research Labs 2013-2015
Postdoctoral Scholar, International Institute of Computer Science, 2015-2016
Research
My research explore the trade-offs between computational effort and accuracy to design algorithms for the efficient solution of large-scale linear-algebraic and machine learning problems. In particular, I have made research contributions in the following areas:
- the design and analysis of scalable machine learning algorithms
- nonlinear and multilinear applications of sketching techniques
sampling for low-rank tensor/matrix approximation and completion
Kernel Methods, Randomized Numerical Linear Algebra, Large-scale Machine Learning, Low-rank approximation and Sketching
Teaching
I regularly teach RPI's introduction to discrete mathematics for computer science for undergraduates, CSCI 2200. See https://www.cs.rpi.edu/~gittea/teaching/spring2023/focs/focs.html for the latest iteration of this course that I have taught.
The upper-level undergraduate and graduate courses I teach concern machine learning. In particular, I regularly teach a course entitled Machine Learning and Optimization, which covers the first order optimization algorithms used in modern ML and introduces students to commonly used architectures in modern ML. See https://www.cs.rpi.edu/~gittea/teaching/spring2023/mlandopt/mlandopt.html for the latest iteration of this course.
I do not have office hours outside of those for the courses I am teaching. Contact me via email to arrange meetings, or alternatively, via Slack if you are an advisee.
If you are my student advisee, you have been invited to a Slack channel that I use for advising. You can contact me through that (preferable), or via email.
I am happy to discuss your course selection given your specific career goals, how to get into undergraduate research, how to plan and execute applications to graduate school, and offer general mentoring and professional and career advice. For most inquiries on paperwork (e.g. registrar forms) or regarding which courses count towards your specific graduation requirements, I will point you towards the Science Hub.
Publications
The following is a selection of recent publications in Scopus. Alex Gittens has 34 indexed publications in the subjects of Computer Science, Mathematics, Engineering.