As part of an initiative known as the Cognitive Campus Project, KIT—working with teams from IBM Software Group Lab Services and IBM Cognitive Business Solutions—implemented an analytics-based advisory service focused on job placement and campus life. Running on a hosted cloud platform, the solution’s fundamental cognitive function is multi-parametric matching performed in two stages. First, the solution creates an extremely detailed, evidence-based profile of each student’s educational identity using a large base of structured and unstructured student data extracted from the institute’s internal systems. With this foundation established, the solution’s second function is to analyze the data to learn the linkages between educational parameters on one side and known career placement outcomes on the other. The practical value of the solution is the ability to sort through these logical connections to offer each student customized advice on which steps to take to achieve a particular career outcome. Suppose, for instance, a student identifies his career placement outcome as living in Silicon Valley and working for a trade firm. The solution’s algorithms—automatically factoring in the student’s background and larger historical patterns—can estimate the value of taking additional English courses, along with recommending additional measures such as taking courses in specific technology subject matter areas. In addition to suggesting the best path to a particular career outcome, the solution has the capability to recommend particular career options that the student may not have been aware of or considered. Using the same basic algorithmic logic, this capability finds the closest matches between the background parameters of the particular undergraduate/graduate student and those of undergraduate/graduate students who have moved on to careers. The solution would present this information as a case study, saying, in effect, “A student with your background followed this particular course and is now working in this particular field.”
KIT expects to increase the number of undergraduate/graduate student applicants drawn by its more sophisticated career placement capabilities. In parallel with this, one organic cultural change expected to unfold over time is an improved alignment between the institute’s curriculum and the demand for skills and academic backgrounds in the workplace. The institute believes that as it gets better at identifying the patterns of success, academics, counselors and curriculum planners will be in a position to make more informed decisions and will therefore be more likely to take proactive steps to keep the curriculum in line with market needs. The same dynamic is also expected to promote closer collaboration and partnership between the institute’s academic staff and the business community.
The KIT solution is game-changing because it uses machine learning technology against large amounts of unstructured data to determine which combination of skills, educational background and activities can maximize a student’s likelihood of succeeding along a particular career path. This ability to provide evidence-based recommendations is likely to improve already good career placement rates for the institute, strengthening its competitive position.
Traditionally, career placement efforts required academic counselors to offer suggestions to students—which skills to improve, which courses to take and which activities to involve themselves in—from their own past experiences. This less-than-objective approach relied on limited inputs and subjective judgment. The new solution takes an entirely objective, fresh and bottom-up look at which factors are most important to success, thereby increasing the likelihood of success and eliminating potential judgment bias in career-related recommendations.
The KIT solution relies on data drawn from its internal systems. Unstructured sources include interview transcripts, class reports, extracurricular activities and course syllabi. Structured data includes course histories, academic records and other elements from the institute’s internal student databases.