1 / 19

Chandu Thekkath Microsoft Research Silicon Valley

Your Research Career. Chandu Thekkath Microsoft Research Silicon Valley. Assumptions and background . You: Graduating or recently graduated Ph.D. Experimental system scientist Research career in industry or academia Me:

hetal
Download Presentation

Chandu Thekkath Microsoft Research Silicon Valley

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Your Research Career Chandu Thekkath Microsoft Research Silicon Valley

  2. Assumptions and background • You: • Graduating or recently graduated Ph.D. • Experimental system scientist • Research career in industry or academia • Me: • SW Engineer (‘83-’89). Ph.D.(‘89-’94) Systems researcher (‘94-present) • Worked in industrial labs (DEC, Compaq, Microsoft) • One year on sabbatical at Stanford • Research manager for 3 years

  3. The talk • Personal, prescriptive, and sometimes contradictory • Based on my experience and many researchers I interviewed: • Fresh Ph.D.s to Turing award winners • 1 year out of school to over 2 decades in the field • Mix of software and hardware systems • Some theoreticians • Both men and women • No guarantee any of this is repeatable

  4. Things that matter • Choice of areas, topics, and projects • Who you collaborate with and how • How you position your research and yourself to the external world • Your management and how you are evaluated • Projects/Ideas • You • Collaborators • Management • External Visibility

  5. Topics, areas, and projects • Picking the problem is usually the hard part • Spades and toothpicks (Roger Needham) • Lofty goals vs. concrete results • Solvable vs. high risk problems • Topics that are fun and meaningful to you • Would you use the system? Would anybody? • Plan to evaluate your system (if not use it) • Problems that annoy you • Find a way to make it less annoying

  6. Topics, areas, and projects • Technology leads systems research • The bleeding edge yields nuggets • Immerse yourself; fortune favours the prepared mind • Inter-disciplinary projects • Conferences outside your area • Be open and broad and experiment with many areas • Research career spans 20-25 years • Change topics. Keeps you alert/inventive • Choose projects that you will learn from

  7. Topics, areas, and projects • Develop good taste • Like wine tasting, the more you do it, the better you get at differentiating good from mediocre • Accept ideas and input from others • Be willing to work on other people’s ideas; they will sometimes work on yours • Avoid the tyranny of the LPU and the paper chase • Focus on learning more than CV building

  8. Your colleagues and collaborators • Good systems research requires collaboration • Multiple researchers with different strengths • People who are smart, inventive, different from you (and nice to hang out with) • Move on if you make a mistake • Don’t obsess over credit; evens out • Don’t antagonize people (we are opinionated and full of ourselves) • Hang around people whose work you admire. Try to become technical friends with them

  9. Your colleagues and collaborators • Neither a minion nor a manager be • Collaborate as a peer (hard if collaborator is senior and in a hierarchy) • Use your colleagues • Advisors, users, critics (c.f. conference PC) • Look for external collaborations • Not always easy: geographical, organizational, political hurdles

  10. Promoting yourself externally • Your research and user community • Publications, talks, participation in conferences/workshops • Making your system available externally • The product division in your company • “Technology transfer”: transferring your knowledge or research artifacts • Funding agencies

  11. Publishing • Number of papers vs. number of citations • Publish only things of which you are proud. Weak papers can mar a strong CV. • Huge difference between writing a paper vs. writing the best/definitive/seminal paper on a topic • 5 papers a year vs. 1 SOSP paper, 5 times in a row. • Avoid boondoggles, however attractive • Ultimately, where you publish will reflect on your CV • Your peers may judge you by your submissions (not just your publications)

  12. Publishing • Write papers that • Teach your readers something • Contain results that can be reproduced • Be generous with credit. E.g., • Don’t get hung up on author order • If you are not presenting at the conference, Use other prestigious venues (e.g., colloquia at top-tier universities)

  13. Publishing • Papers get rejected for many reasons. Don’t lose confidence and self-esteem • Find venues to talk about rejected work if you think the work is good. • Ask for advice from your more experienced colleagues • Fix the shortcomings of the research. • Take your time, re-evaluate your decision to republish • Be a conscientious reviewer/PC member • Word gets around if you are a sloppy/indifferent PC member

  14. Technology transfer • Don’t confuse research, advanced development, and product development • You may do all three at times, but be aware of what you are doing • Product folks have a different mindset • Respect them; they are smart but they march to a different beat. • Don’t expect your research system to be immediately useful • Gestation period can be a few years • Develop a research pipeline

  15. Technology transfer • Use product development to learn about real problems • Real usage reveals real problems • Move between research and product development (gasp!) • Work as an advisor or architect • Attend design reviews

  16. Management & environment • Management style matters in research • Top down vs. bottom up can make a huge difference • Research is best done from the trenches rather than from the hilltops • Make sure you know how research is actually done in the organization • “Research” or “Researcher” may not mean what you think • Do you have freedom to pursue your research agenda? • If not, how much of it is directed by group lead/manager • Can you live with this?

  17. Management & environment • Is research your primary responsibility? • Or is it done in your spare time • Is it done only for a fraction of your time • Will you be rewarded for doing only research? • No product impact for an extended time? • Do projects have to be “approved” or “funded” • By somebody (e.g., your manager) • By some group (e.g., a product division) • Do you have enough resources for • Travel and equipment • Visitors/collaborators/interns

  18. Odds and ends • Living your life and being a researcher at the same time • What fraction of your social/family life are you willing to give up. • Pick a balance that works for you. • Being happy is more important than being a star researcher • Know when to give up • If you are not having fun being a researcher, do something else. • Many fulfilling non-research careers available for a Ph.D. in CS

More Related