Desktop version

Home arrow Computer Science

  • Increase font
  • Decrease font

<<   CONTENTS   >>

A View from the Ground

The following is a view of working with universities from the perspective of a senior data scientist. If you are new to collaborations, you may find this account useful when considering your own needs, or if you have worked with universities in the past, you may find these reflections a good comparator.

Table 2.4 Faculty Rank, Tenure Status, and Incentives for Corporate Collaborations

Academic Rank

Tenure Status

Faculty Incentives for Collaboration

Lecturer/Professor of the Practice or Clinical Faculty

Non-Tenure Track (often no PhD)

No research or publication requirements. Primary incentive for collaboration is in-class speaking and capstone project sponsorship. These faculty are almost exclusively aligned with undergraduate programs ("Clinical Faculty" or "Professor of the Practice" may be aligned with the Business School, i.e., the MBA Program).

Assistant Professor

Non-Tenure Track (PhD)

No research or publication requirements. Primary incentive for collaboration is in-class speaking and capstone project sponsorship. These faculty are almost exclusively aligned with undergraduate programs.

Assistant Professor

Tenure Track

Heavy emphasis on research and publication. Sponsored research (with the opportunity to publish) would be the strongest incentive. These faculty may work with students at the undergraduate or master's levels.




Combination of research, innovation, consulting, community engagement, and student capstone project sponsorship. If the faculty is working on a federally funded grant, they may not be available to work with any other partners. These faculty may work with students at any level. If they have a strong research orientation, they will be aligned with a doctoral program.



Consulting. Potentially sponsored research. Community engagement. These faculty may work with students at any level.

Khalifeh Al-Jadda, Director of Data Science, The Home Depot

Data science as an emerging field has been one of the most challenging areas of recruitment for all companies. However, due to the significant impact data science teams have made in their companies and the great ROI in this area in many companies, most business leaders and executives see the value in investing in data science. Therefore, many companies turn to universities as an important source of talented recruits for data science teams. The unique thing about data science is how close it brings the industry and research communities to work with each other since research skills are essential in any successful data science organization. That said, any successful data science team in industry relies on hiring graduate students from research labs that actively work on data science areas like machine learning, computer vision, deep learning, NLP, etc.

During my tenure as a data science leader in different companies, I have established different forms of collaboration with universities, which helped me to achieve great things that cannot be accomplished without those partnerships.

These forms of collaboration include serving on advisory board of data science programs at universities to establishing internship programs focused on research and development. Below, I highlight my experience in different forms of collaboration with universities.


One of the most successful hiring strategies in data science is the internship.

I personally started my career in data science as an intern at CareerBuilder which gave me exposure to solving real-life problems and working on large scale datasets at the enterprise level. After I become a data science leader and hiring manager, I started to leverage the internship as an exploration arm of my teams. We hire PhD/ MSc students every semester based on their research background then we assign them a research/discovery project to work on for three months while one of the fulltime data scientists within the team mentors them. Those interns in turn contribute significantly to our exploration effort by building POCs during their internship using cutting-edge techniques and new models that we have yet to explore as a team due to our busy schedule maintaining and optimizing existing models that power our production environment. On the other hand, such internship programs contribute significantly to our portfolio of patents and research papers which play an important role in attracting top talents to join our teams.

Here is a list of benefits that I believe a strong data science internship program can bring to any data science leader:

  • 1. Discover new talented recruits for the full-time opportunities and avoid any false-positive hiring.
  • 2. Explore new models and techniques to solve challenging problems.
  • 3. Enrich the team's portfolio of publications and patents.
  • 4. Reduce the cost of research projects incurred by contracting with outside firms.

Research Collaboration

Another interesting form of collaboration is in Research. In this form of collaboration, we find a research lab that works on one of the challenging problems that we would like to solve for our company. Therefore, we offer financial support to that research lab either as a gift or as a scholarship to one or two of the graduate students of that research lab then we start working with the professor and graduate students in that research lab on our own dataset so as to leverage their techniques and algorithms on our datasets to solve one of our challenging problems. This form of collaboration is usually a long-term investment where we work with the research lab for six months to a year or more, and solve challenging problems at a large scale, which cannot be solved using a single intern over a semester (e.g., building knowledge graphs or deep learning frameworks). However, the data science leaders must always be careful since such collaboration will require companies to compromise their IP rules and be willing to give up (or at least share) any patents that come out of such collaboration. Moreover, there is a risk that such collaboration may not have any direct impact on the business since some research labs tend to focus on the research side of the work instead of the applicability and scalability of their solutions.

The pros and cons of a research collaboration:

  • 1. Produces high quality research which usually contributes significantly to the organization's portfolio of industry thought leadership and patents.
  • 2. Solves challenging problems at a large scale.
  • 3. Allows flexibility since the research lab will have access to the company's data and they will work in the lab according to their own schedule.
  • 4. Saves the cost of hosting the researchers in the company's office.
  • 5. Risks not having a direct impact at the end of the contract.
  • 6. Requires companies to compromise their IP rules and give up some patent rights.

Advisory Board Membership

I served on the Board of Directors of the Analytics and Data Science Institute at Kennesaw State University. This unique experience gave me a chance to have a closer look at the challenges that face data science educators. Before joining that advisory board, I wasn’t aware of the lack of funding that such programs face to attract faculty members with the right research background to keep up with the industry needs. On the other hand, as an industry leader I was able to share industry expectations from the graduates of such programs, what are the core skills that we care about and the research areas in highest demand. Moreover, being on that board of directors gave me an opportunity to work closely with the professors and instructors to share some of the industry use cases and recent work so the students can prepare themselves for the transition from research labs to industry.

Overall, serving on an advisory board:

  • 1. Makes you more keenly aware of the challenges that face the data science programs at universities and can try to help solve some of those challenges.
  • 2. Bridges the gap between industry and academia by sharing the expectations and core skills needed in the industry to help those who run the data science programs in universities focus on those skills.
  • 3. Helps students to understand the difference between the research projects in their labs and the practical data science projects in companies.

Our Summary Checklist for Working with Universities

For any manager of an analytical organization interested in approaching a university for the purposes of recruiting talent, establishing a pipeline for talent, sponsoring a capstone course or project, collaborating in research, or executive education, we have provided a checklist to get started. As you think about what you want out of a university engagement, we have provided more specific checklists in the succeeding chapters. As a starting point, consider

/ Be clear on what you and your organization are expecting to get out of the relationship.

/ Why do you want to partner with this university? Have the graduates from this university been successful at your organization? Are there particular faculty members that produce research or teach courses that are well aligned with your organizational objectives?

/ Develop an understanding of why the university and/or faculty would want to collaborate with you and your organization. Remember that different types of faculty have different incentive systems depending on where they are in their career. They need to have a reason to return your call. Funding is important, but not sufficient.

/ Understand that one size does not fit all. Companies with a successful history of university collaborations typically have a portfolio of university relationships.

/ Consider your appetite for sharing intellectual property and publication of research findings. For research faculty and most pre-tenure/tenure-track faculty this will not only be a priority but if there is no appetite at your organization for sharing intellectual property, the faculty will be dis-incented from working with you.

/ If your organizational priorities are particularly related to hiring entry level talent, consider partnering with strong undergraduate and regional institutions, which may be easier to “navigate”. In addition, there are broader latent “community-based” benefits to hiring locally. Alternatively, if your priorities are related to research, consider reaching out to large research universities with centers and institutes.


  • 1. Industry-University Cooperative Research Centers (IUCRC), National Science Foundation, Accessed June 10, 2020.
  • 2. Tire Carnegie Classifications of Institutions of Higher Learning, https:// Accessed April 20, 2020.
  • 3. Colleges that Change Lives, Accessed June
  • 10, 2020.
  • 4. Association to Advance Collegiate Business Schools, Accessed June 3, 2020.
  • 5. Accreditation Board for Engineering and Technology, Accessed June 3, 2020.
  • 6. Drexel University, Accessed April 13, 2020.
<<   CONTENTS   >>

Related topics