Kern Community College District partners with AWS to build AI/ML-enabled Guided Pathways tool

Students attend community colleges for multiple reasons. Sometimes they drop out—for reasons that aren’t always entirely clear. Now, one of the largest community college districts in the country is determined to not only get a clearer picture but do something about it. Kern Community College District (KCCD) district in California is partnering with Amazon Web Services (AWS) to design and build a student-centered course planning tool to support the district’s Guided Pathways model.

By using artificial intelligence and machine learning (AI/ML), KCCD and AWS are optimizing student course-taking patterns to decrease time to degree, improve enrollment forecasting, and focus limited staff resources where and when they are needed most. The result? KCCD leadership and practitioners will gain dynamic insights into successful course-taking patterns they can use to provide focused interventions and understand course-offering needs and systemic blockers. Most important, students will be provided with evidence-based enrollment guidance that can help them finish their degree faster, with fewer excess credits, and at a lower cost.

A commitment to equity for all students

Covering 24,800 square miles across five counties, KCCD serves a diverse student body population of 40,000. The district is composed of three colleges: Bakersfield, Cerro Coso, and Porterville College, with the district’s reach extending deep into the communities they serve to community education centers and remote learners.

Despite its size and geographical breadth, KCCD maintains a commitment to an equitable system of education for every learner. “Older, legacy technology won’t help students reach their goals and dreams,” said Gary Moser, KCCD chief information officer. “Only the latest technology will do that.” But accessing top technology is difficult given the constraints of budget, which, for a community college system, notes Moser, “is limited compared to our four-year counterparts.”

Moser knew that using specific new technologies—like AI/ML—could help KCCD meet student needs while also maximizing budget and impact. “Effective technology also makes us more efficient to support staff and faculty, and of course, to help manage costs,” said Moser.

AI has the ability to augment staffing that has been challenged with the staff-to-student ratios, freeing them up to focus on working with the students who need them when they need them—a win-win-win for staff, students, and cost management. 

KCCD’s partnership with AWS addresses a growing need

Students attend community colleges with different goals. For some, it’s to earn a degree or credential and enter the workforce. For others, the end goal is a bachelor’s degree at a four-year college. The contemporary learner profile is more diverse than ever, and whatever their motivations, they all share one thing in common: nobody wants to leave school without completing the degree, certificate, or transfer requirements they sought to achieve.

When KCCD and AWS set out to design the Guided Pathways tool, it was to answer a fundamental question: how can we use AI/ML to optimize advising and counseling resources and help students reach their graduation goals as quickly and with the fewest excess credits possible?

To help answer this question, KCCD advisors and counselors review a number of different data points, including student goals, transcripted courses and grades, degree requirements, and financial aid status. While this information existed within the KCCD system, it was scattered across multiple siloed data systems. With average caseloads in the thousands, advisors and counselors needed the information from the disparate systems to be consumed, analyzed, and shared in a way that was timely and easy to access and understand.

Guided Pathways provides near real-time insights into student behaviors

The KCCD Guided Pathways tool was built for counselors and advisors to automate real-time insights into student enrollment behaviors. This is so as they guide their students toward their academic and career goals, they know right away when a student takes courses that are not optimal for their identified program of study. This dynamic notification allows for advisors and counselors to intervene before the cost of enrolling in the wrong classes impacts a student’s pace or progression or, worse yet, derails them altogether.

In addition to decreasing students’ time to degree, the tool offers a way to optimize financial aid by decreasing excess credits, lowering the final cost of their degree, and assuring students are maintaining federally mandated satisfactory academic progress. Getting off track with financial aid has impacts on future access to financial aid. It doesn’t just delay degree completion and increase costs, students can run out of financial aid eligibility altogether.

Collaboration with AWS results in innovation

For Moser, working with AWS has been a true collaboration. “They’re a worldwide leader in creating solutions,” he said. “They’re there to work with you, not just sell you something.”

KCCD’s deputy chief information officer, David Barnett, together with the Institutional Research Department, worked directly with AWS solutions architects to pull in the necessary data from the district’s enterprise resource planning (ERP) system. Using AWS Cloud services and the AWS ML tools, KCCD and AWS worked together to build out a proof of concept (POC) of the solution. With the initial phase of the POC complete, the solution has entered the next critical stage: active testing with KCCD counselors and advisors, the first set of stakeholders.

Gaining critical insights into student enrollment patterns

Initial testing reviews have been positive, with one reviewer sharing that this would allow them to focus their efforts and change their intervention strategy. Now that initial validation is complete, the team will present the results to leadership across the three colleges.

“It’s important that everyone feel both confident and comfortable with the system,” said Moser.

Although metrics are still being developed, data will likely include the amount of financial aid saved, excess credits eliminated, number and type of changes to their program of study students can make without financial or degree progress impact, and retention and graduation rates.

Among the insights the KCCD team and Moser expect to gain from the Guided Pathways solution are:

• why some students stay enrolled for more than the typical two years that full-time students take to complete coursework
• what courses students enroll in that were not on their degree plan and why
• how the model could be applied to other student data points (co-curricular engagement patterns)

“And this is where Guided Pathways will help, by eliminating the confusion and delays that lead to students giving up,” said Moser.

Using AI/ML to guide the future of education in California

While the solution is still in its POC stage, Moser is already excited about the ways AI/ML is poised to improve education across the entire state of California. Moving forward, the district would like to expand use of the Guided Pathways tool, eventually offering it statewide. Moser can even envision the partnership scaling beyond that.

“We aren’t just piloting a Guided Pathways tool here,” he said. “We’re developing a process for AI-solution development in general.”

Colleges and universities around the globe rely on the AWS Cloud for tools such as financial aid tracking and intelligent document processing. Learn more about AWS Cloud for Higher Education.