I wanted to be a teacher ever since  my seventh grade geography instructor, Mr. Fortunato, let me teach the class for 15 minutes! We lived in New Jersey and before we packed up and moved to Michigan later that year, I made my parents  take me to see the Montclair State Teachers College campus.

I went to high school north of Detroit and then to the University of Michigan, College of Engineering.  I started out in Naval Architecture and then switched to Meteorology and Oceanography. But by the end of my sophomore year it was clear that I really loved mathematics and that the only way I could get enough was to major in the subject. The College had a program in  Applied Mathematics and I affiliated.

In my junior year I took my first course in Linear Algebra from my advisor, Professor Jack Goldberg–it was great. In my senior year I took a 2-semester sequence in numerical analysis from Cleve Moler. We used Cleve’s book with Forsythe on linear equation solving and I bought a copy of Wilkinson’s Algebraic Eigenvalue Problem. That was my start in matrix computations where I have worked ever since. Cleve was a great teacher and I inherited from him all my perspectives on computing and mathematics. I remember going to an  evening “Math Club” seminar where he gave probably the first-ever talk on Matlab.

I supported myself as a teaching assistant and by the time I left U Michigan I had 4 solid years teaching freshman-sophomore calculus. At Michigan, the TAs ran the whole calculus “show”. It was a great experience, ranking second only to my 15 minutes in Mr. Fortunato’s class!

I received my Ph.D.  in 1973 studying under the direction of Cleve Moler. My dissertation was on a matrix problem known as the generalized singular value decomposition (gsvd) problem. To carry out the calculations  I developed an algorithm which  generalized just about every known eigenvalue/singular value procedure around. There weren’t too many realistic gsvd applications back then, but later on in the 80’s some critical star wars applications had  gsvd solutions. Reasonable people can differ on this, but I like to think that it was (the threat of)  real-time, matrix-based signal processing that brought down the Berlin Wall! 

After I graduated I spent a year and a half at the University of Manchester as a post-doctoral fellow in the Department of Mathematics. It remains one of the high points of my academic career. Nothing compares to the freedom and excitement of a post-doc.

Around about that time I met Gene Golub and we had some great collaborations. In the summer of 1977 we co-taught a workshop on matrix computations at Johns Hopkins and from that developed our book on the subject.

I have always worked in the area of matrix computations and have written papers on various eigenvalue, least squares, and linear equation problems.   To me, the matrix problems that arise in signal processing and control theory are particularly interesting. Prior to becoming Professor Emeritus in 2016 I was a researcher in the area of computational multilinear algebra.

I have written six textbooks: Insight Through Computing–A Matlab Introduction to Computational Science and Engineering (with D. Fan),, Matrix Computations, Fourth Edition (with G.H. Golub), Handbook for Matrix Computations (with T. Coleman), Computational Frameworks for the Fast Fourier Transform, Introduction to Computational Science and Mathematics, and  Introduction to Scientific Computation–A Matrix/Vector Approach Using Matlab.

I  have been a Professor of Computer Science at Cornell since 1975. It’s the perfect place to work for someone like me. I can’t walk across the campus   without bumping into an interesting matrix problem.  And the Department has given me the time and intellectual space to develop some courses that are truly fascinating–to me at least!

I’ve have had several major administrative positions since I joined the faculty. From 1982 to 1987 I ran the CS graduate program which (at the time) enrolled about 80 PhD students. It was a lot of work but very rewarding. I especially enjoyed the admissions process and watching the students I admitted go through the program and on to great careers. I also learned a lot about the psychology of getting a PhD and just how much of it depends on self-esteem.

From 1994 to 1998 and 1999 to 2003  I was the Director of the Undergraduate Studies. Here, the number of students was much larger–there were about 500 CS majors to track in those days. The challenge was  to make our program among the strongest in the country and to build a reputation that will help attract the best high-schoolers to Cornell. This meant keeping the curriculum up-to-date and taking steps to ensure that the requirements are consistent with the  aims of liberal education.

From July 1999 to June 2006 I was Chair of CS. It was an exciting period with the creation of CIS in 1999. Few things in academic administration compare to the hiring  of fresh PhDs into assistant professorships.

From 2013-2016 I  ran the CS Master of Engineering program. I also found that to be extremely interesting given the heightened profile of entrepreneurship within academic CS and (relatedly) the creation of Cornell Tech in NYC.

I retired from CS on June 30, 2016 and became Cornell’s Dean of Faculty (DoF) the day after on July 1, 2016, becoming the first emeritus faculty member ever to be elected to that position. I was the  DoF for five years and was perfectly prepared to take on all the challenges thanks to my experiences as a CS faculty member. Those experiences included being Department Chair for 7 years, Director of Graduate Studies for 5 years, Director of Undergraduate Studies for 8 years, and Director of Masters Studies for 3 years. The CS environment  is collegial and a great place to learn about communication, transparency, and the connections between research and teaching. For me, creating and teaching CS 101 (“Programming for Poets”) in the late 1970s clarified my thinking about the connections between of computer science and liberal education. On the research side, I was lucky to work in the numerical linear algebra area–its centrality taught me all about the relationship between math and CS and how both fields shape engineering and science.