Subscribe to DSC Newsletter

You may even call it post-doctorship, as the level is beyond the traditional PhD degree. It is not a degree, not competing with university programs, but instead, akin to a fellowship or apprenticeship to learn doing state-of-the-art applied research, discover ground-breaking results or applications, and translate your discoveries into seminal material suitable for a broad audience. It is intended for professionals with substantial experience, perhaps to people who already have a PhD in a different field. It is mentored by well connected, world-class recognized scientists (not necessarily affiliated with a university) with broad domain of expertise in many environments. The focus is on real-world problems and applications to help you get a high-level position in the industry or as an independent researcher.

The idea to create such a program stems from the fact that a PhD degree is sometimes perceived negatively when looking for an Industry job, and some candidates hide it in their resume. This is because a PhD degree is supposed to prepare you for Academic research, but due to the large supply of PhD's and a shrinking job market for tenured positions in Academia, many are left in precarious situations.

Source for picture: Women in data science

This doctorship is designed to make you a leading scientist in the Industry, to become for instance, an executive data scientist or chief scientist. It does not involve writing and defending a thesis, nor publishing esoteric papers in scientific journals read by few, but instead to quickly disseminate your research, explained in simple English, to a broad audience of practitioners.  For instance, possibly to attract VC funding if you want to create a start-up out of it. The standards are by no means inferior to that of traditional PhD programs, they are just very different. The length of this "mentorship" could be as short as two years; it could be carried out part-time, remotely, while having a full time job at the same time.

Features of the Proposed Doctorship

  • Short duration (2 years)
  • Publication in niche media outlets (like DSC), not in scientific journals
  • No teaching load
  • Done remotely, part-time
  • Cross-disciplinary research
  • Not attached to a university program
  • No thesis, no defense
  • Focus on Industry problems
  • Research not influenced by grant money or politics
  • Candidate gets an apprenticeship in the corporate world, related to her research
  • Not a degree
  • Not meant as a substitute to PhD programs

Example of Research 

I have one example in mind. This is what I would offer if I can find the time to start such an endeavor. The research in question is at the intersection of data science, dynamical systems, stochastic processes, computer science, and number theory, with applications to cryptography, Fintech, Blockchain, security, and high performance / high precision computing. Side projects could include the design of continuous random number generators, replacing standard statistical tests and p-value by better tools, or proving that the digits of some mathematical constants, are randomly distributed (this would be a fundamental result.) See here for details. I believe this research could lead to ground-breaking discoveries and nice applications.

I don't even have a PhD in data science, such PhD's did not exist back then. Instead my PhD was in computational statistics. Do I qualify to run such a program? Of course it is not for me to answer that question. You can check my career path here and judge by yourself.

The problem is, how many qualified experts are willing to take the challenge to offer this type of mentorship? How many professionals are willing to join such a program? Are there any companies interested in joining as partners?

Related article

For related articles from the same author, click here or visit www.VincentGranville.com. Follow me on LinkedIn.

DSC Resources

Views: 21049

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Dr. Austin Umezurike on March 28, 2018 at 9:41am

This is an interesting concept. It should be packaged as  Data Science Post-Doc fellowship.

The duration should be about six months to sustain interests and prepare candidates for real world challenges

Thank you,

Comment by Vincent Granville on March 26, 2018 at 6:05pm

Hi Robert,

To answer your question, I am not even sure that I am the right fit for this type of project. The reason being that in my case, I would incorporate advanced number theory in the research, because of my own interests in this topic. It would still be accessible to neophytes in the field, with spectacular results expected to make the successful candidate being remembered for hundreds of years (e.g, proving that the digits of Pi or other math constant behave randomly in some number representation systems -- I am getting close to proving this one.) 

However, the focus in my case is on applications (Blockchain, Fintech, cryptography, security, business analytics.) I built a framework that I believe will be valuable both to number theorists and applied data scientists. Even the theoretical component relies heavily on HPC and data science, for discovery purposes. This is the kind of research that, even though it has ground-breaking potential, would have a hard time being funded by a grant.

I expect that other mentors interested in this initiative, would not have that sort of theoretical component in their mind (and even in my case, it is not about writing esoteric papers, but instead seminal discoveries that the layman can understand and asses its value, with obvious applications to real world problems.)

I have reached a point in my career where, I believe that I am about to publish (here) the best of what I have ever done in my mathematical / data science life, both applied and theoretical. My expectation is that any mentor interested in this initiative would be in a similar situation, with an interest to work with the best apprentices to deliver stellar research results, help them create their own company, or help them join a top company as a Chief Scientist.

Vincent

PS: Doctorship is an archaic word for doctorate, no longer used. It was quite popular three centuries ago. PhD stands for Doctor of Philosophy, though nowadays there is not much philosophy involved.  

Comment by Robert Nisbet on March 26, 2018 at 2:22pm

My thoughts on this proposal are conditioned largely by my experience of teaching online data science courses in the Data Science and Predictive Analytics certificate programs at the University of California, Irvine.  These certificate programs are focused primarily on aspects of the practice of Data Science, not the theoretical aspects.  These programs have been hugely successful since 2010.  Now, there are many similar certificate programs available from other colleges and universities.  Any doctorate program formatted as you describe should plug into that trend, not the academic research-oriented world.  The program should focus on practical problems in the process of data science, not those theoretical aspects of such great interest in academia.  Specifically, it should focus on more efficient methods for data integration, data preparation (my specialty), integration of automated processing with manual processing, and deployment of models company-wide.  James Taylor wrote his book, "Smart (Enough) Systems" upon the premise that it may take only 10% of the modeling time to produce a model that is 90% as accurate as the most enhanced model can produce, and that is good enough for most business applications.  Time is money in business, and academic refinements may not be worth the time and money to achieve them.  This approach underlies all three of my books in Data Science.  My background is in ecosystem modeling.  I began to apply the ecosystematic approach to business systems at AT&T in 1994, followed by NCR - Teradata..  Now, many large companies have climbed on board this wagon (e.g. Apple, IBM, SAP, Oracle).  One  focus of such a doctorate program might be to design model deployment systems that integrate all business departments/divisions of a company.  That is the way we did it with the compartments of complex forest ecosystems at UC-Santa Barbara.  If you would like to pursue such a practically-oriented doctorate program, I might be interested in working with you.

Bob Nisbet, PhD

Instructor, Data Science Certificate Program

Division of Continuing Education

University of California, Irvine.

Comment by Vincent Granville on March 26, 2018 at 11:43am

Dan, I agree with most of what you wrote, but at the same time, I do not see this doctorship as being delivered by a university. A university could certainly offer a PhD in data science (and they could sell it as an accredited degree) but the doctorship in question, (If I was a mentor) would have a very strong theoretical component, deeper than what I worked on when I earned my own PhD. Along with modern applications. The goal is very different and definitely business-oriented (and also allowing people to work on projects that no grant will ever fund), and the means are very different  too (part-time, remotely, no teaching load, no defense -- instead you spend all your time on productive research, not unlike some post-doc programs, which is why it is much shorter.) 

Comment by Dan Butorovich on March 26, 2018 at 11:23am

I somewhat agree with Patrick on this issue but I also somewhat disagree. I agree that Data Science as it exists now is an almost entirely practical field that is not academically focused but rather focused on industry and solving real world issues. However, that said, I believe that there is room in academia for data science in the same sense that we have Doctorates in Business Management. Do you need a PhD in Business to run a company? No, but it helps solidify the field *in academia*, and structure the curricula when we generate MS level MBA's who work on real world problems.

The PhD, as a doctorate in philosophy, is mostly concerned with the theoretical. The field of Data Science could use people working at that level on deep level theoretical problems that plague the field right now and exploring new ways to understand data as a science. A PhD would have less interest or use in the often nuts-and- bolts work of the typical data scientist (if there even is such a thing), unless it tied to a deeper theoretical issue. 

In other words, its a good idea as long as it is understood that it is not something needed to work in the field.

Comment by Vincent Granville on March 26, 2018 at 11:18am

I am exploring whether great partners could join (Google, Facebook, Intel, NSA, IBM, Harvard, etc.) I am hopeful that it can get started with great mentors and great partners. If I were going to be one of these mentors, and given the type of research that I am working on, it would definitely be intended to professionals who have developed provable maturity in both mathematics and data science, possibly candidates who already have a  PhD in mathematics or computer science.  To put it differently, it is the exact opposite of what a data camp or Coursera training is. This kind of training is extremely valuable for beginners, but this is not what I am offering. What I have in mind is a bit similar to an MBA from Harvard, with two differences: Focus is on research more than on business, and while the value of an Harvard MBA is the connections that you earn, in this doctorship, connecting with leaders is not the main purpose of the program, though a significant part of it nevertheless.

Comment by E. Amaka Nwankwo-Igomu on March 26, 2018 at 7:00am

Interesting and interested. :)

Comment by Ariel Gamino on March 26, 2018 at 6:56am

This is a great idea. Beam me up!

Comment by Krishna Baskar on March 26, 2018 at 6:17am
Dear Vincent,

I follow you in linkedIn and enjoy reading your articles rehulsrlt. and also a member here for long time.
I am a practicing Data Scientist.
Completed 2 certifications so far in order to acquire knowledge in this field.
1) Data Science Certificate (4 course) from Harvard Extension.
2) Data Science Specialization (10 course) from Coursers (John Hopkins)

I am a strong believer in data science and would like to take-up your Doctorship. Kindly let me know how to get admitted.

Thank you.
n
Comment by Robson Chiambiro on March 26, 2018 at 4:18am

Very interesting 

Videos

  • Add Videos
  • View All

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service