Join us

A good starting point is my research page. Current themes include model compression and quantization for large models, distributed inference under communication constraints, compression with locality constraints, and adversarial communication. You do not need to have worked on these before, but you should find at least one of them genuinely interesting enough to read a paper about.

Regardless of level, to do good research, you need to:

• be comfortable with probability, linear algebra / matrix theory, and programming in Python and/or C/C++;
• enjoy thinking carefully about a problem and formulating these carefully before coding, and be honest about what they do and don't understand;
• be persistent: research means being stuck often, and continuing anyway;
• be able to work independently while communicating clearly and regularly.
• Prior exposure to research is not a prerequisite. Evidence that you can learn hard material on your own, and that you finish what you start, matters far more.

New students usually start on a well-defined problem with a clear first milestone, so you build momentum early; over time you take increasing ownership of your own research direction. We meet one-on-one (typically) once a week. I am hands-on with students who are new to research and deliberately step back as you become more independent. I expect honesty about negative results, and hiding what didn't work helps no one.

On AI tools: you are welcome (and to some extent, encouraged) to use them to help with coding, literature search, and clarifying concepts, but you are expected to understand and verify everything that goes into our work. The goal is for you to become a stronger researcher, not to outsource the thinking.

• PhD students: Admission to the PhD program is through the IITH / EE department process. If you would like to work with me specifically, email me (details below) after or alongside your application.
• MTech/MSc students at IITH: If you are (or will be) an MTech/MSc student at IITH and want to do your thesis/project with the group, email me with the details below. Taking one of my courses first is a good way for both of us to see if it's a fit.
• BTech students at IITH: I regularly supervise BTech projects. If you have taken a relevant course or are willing to put in the background work, please let me know. However, it is helpful if you can commit at least 6 months to 1 year into working on a problem.
• Interns from other institutions: Please read this before emailing. I receive a very large number of internship requests and cannot respond to all of them. I consider a small number of interns each year, and only when there is a clear research fit and, usually, a funding or fellowship route (e.g. SURE internships during the summer). Emails that follow the format below are the ones I am able to read and reply to. Currently, I do not have any other funding for supporting internships

To help me give you a real reply, please send a single email to shashankvatedka<at>ee.iith.ac.in with the subject line [Prospective <PhD/MTech/BTech/Intern>] Your Name, containing:

• Who you are: your program, year, and institution.
• One paper or topic of mine you read, and one or two sentences on what interested you or a question it raised. (It tells me you actually looked.)
• Your background: relevant courses, projects, or skills; attach your CV and transcript as PDFs.
• What you're hoping to do: rough duration and, for interns, your funding situation.
• Please don't send mass/templated emails. These are generally easy to catch, and I won't reply to them. A short, specific, honest email framed in your own words always gets priority.