Current StudentsIf you are a current Caltech student and wish to meet with me, please clearly state WHAT you want to discuss and a range of AVAILABLE TIMES that you can meet. This will reduce the amount of back-and-forth emailing to nail down a meeting time. I typically take meetings in my office (303 Annenberg).
If you just need your add/drop card signed, my administrative assistant, Diane Goodfellow, can take care of that (provided it's not an overload or underload situation).
I have a limited number of research projects suitable for motivated undergraduates (via CS 80 or CS 81). If you are interested in such a project, please email me your resume and unofficial transcript. It is typically easier for me to work with you if you have already taken CS 165 or CS 159, both of which are research-grade machine learning courses.
If you are a current student and are interested in TAing for me, please sign up for the internal TA signup spreadsheet. I will recruit from that spreadsheet in the preceding months before the course starts.
Prospective StudentsUnfortunately, I receive an overwhelming amount of email from prospective students. I'm really sorry if I don't respond to your email. However, since many prospective students ask the same questions, here is a FAQ:
- Will you be accepting new students? Yes, I am always considering accepting new PhD students, especially into Caltech's new PhD program in Computing and Mathematical Sciences, or the PhD program in Computation & Neural Systems.
- Is there anything special I should do when I apply? Yes, please mention me specifically in your research statement when you submit your application.
- What are you currently working on? My research interests are described here. You can read or skim some of my recent publications to get a sense of the technical and application flavor of my recent projects. Most research projects I'm interested in initiating are motivated by recent research that I've done.
- How will I get funded? All PhD students in the CMS and CNS programs will be guaranteed funding for their first few years (e.g., through teaching and research assistantships), and in practice funding is basically never an issue.
- Can I be co-advised? Yes, co-advising is common at Caltech. Perhaps 10-15% of PhD students in CMS are co-advised. Most PhD students do some collaborative work with other professors, even if there is not an official co-advising relationship.
- How does Caltech differ from other schools? Caltech is a very small school, with only about 300 faculty across the entire institute. The Computer Science presence is accordingly also very small. As such, we tend to attract students who have both strong foundations, as well as the potential to engage in fundamental yet cross-cutting research. For example, it is quite normal for someone to work in Machine Learning and have a very concrete connection to applications in Biology. See also Adam Wierman's RSRG Blog post describing our CMS PhD program.
- What do you look for in PhD applicants? Typically, a strong PhD applicant has demonstrated some combination of sustained intellectual curiosity, mathematical maturity, programming savvy, a knack for finding creative solutions, the ability to push forward projects, resolve under duress, clarity of thought & exposition, and an appreciation for elegant insights. Obviously, not every first-year student will be strong in all aspects to start with, but a strong machine learning PhD student will have mastered all these traits by graduation.
- What is the application process like at Caltech? At Caltech, we invite select students to an interview visit day before making final admissions decisions. While not strictly necessary, we are generally much more comfortable admitting students that we've had a chance to talk to and interact with. The interview visit day is typically in early to mid February, and is also an opportunity for students to get a feel for the faculty and the school.
Prospective PostdocsI am open to taking on postdocs for all of my active research areas (you can see my recent publications to get a sense of what I'm currently interested in). To apply, please submit an application here. Please also mention my name in the application as well as your interest in the DOLCIT research group, and send me an email telling me you've applied (although I may not be able to respond to your email, it will help me keep track of applicants).
It is common for postdocs to be co-sponsored by multiple professors, and it is possible for me to co-sponsor a postdoc with any other faculty in CMI. However, certain professors share more research interests with me than others. The short-list includes:
Industry ConsultingI do a little bit of industry consulting. However, I receive many more consulting requests than I can handle, so I apologize in advance if I do not respond to your request.