I am a 4th year PhD student in Computer Science. For my PhD thesis I have worked on problems in Stochastic systems. Currently, I have a publication (accepted) in a Tier 1 Applied Math. journal and another under review in another Tier 1 Applied Math. journal. Also, one work is in progress. Now, I have started thinking about applying for Postdocs. I heard that 3 letter of recommendation is needed. For my PhD thesis I have worked only with my advisor (I am the lead author in all my papers). Hence I am confused about whom to ask for letter of recommendation other than my advisor. One option is my course instructors (I got good grades in all my courses). Could anybody give me any suggestion?
For getting PhD one's PhD thesis is going to be scientifically evaluated and reviewed. It depends on specific regulations but usually such PhD reviewers are people outside of your department/university. So,
together with your PhD advisor these reviewers are the best folks to deliver recommendations (as they are truly capable of evaluating your PhD work & you). The other guys who can be great references for you are those who reviewed your published works (though not explicitly disclosed it might be not hard to figure out that; your advisor might actually know them ;) or those who are familiar with your research by other means (e.g. cited it, referred to it in their talsk/seminars/etc). Lastly, please do challenge your PhD advisor with this issue: in fact, it is one of his/her responsibilities towards his/her PhD students; particularly, it is a PhD advisor who 'orchestrates' a PhD defence process, including e.g. finding thesis reviewers, etc.
p.s. At PhD-->postdoc stage, a course instructor is, in general, not a great person to provide a letter of recommendation. The reason is simple: your potential postdoc bosses would primarily like to know about your research work, capabilities, potential, etc. Your coursework performance is rather irrelevant. Of course, not all course instructors are bad: the one behind a so called PhD-level research-oriented seminar course might be a great reference or it might be more or less 'regular' course but taught by a well-known guy in the area of your postdoc research interests.