California Community Colleges Technology Center (CCCTC) CTO, Lou Delzompo presented today at the Amazon Web Services (AWS) Global Breakfast in Las Vegas.  His presentation showcased work being done in California Higher Education utilizing the AWS cloud, and highlighted recent machine learning work the CCCTC has been focused on this past year.

CCCTC serves one of the largest community college systems in the world. With 114 colleges across the state of California, nearly 4 million students from all over the world enroll to build their career skills, prepare for transfer to four-year colleges and universities, or simply enrich their lives through learning.

“The ‘North Star’ from our Chancellor is to increase the number of successful graduates each year. This is better ROI for taxpayers and helps with the economic development of our citizens,” said Lou. The California Community College (CCC) system aims to increase the number of students who receive degrees or other credentials by 20 percent in the next five years.

The first challenge CCCTC faced was to improve their application process. CCCApply is the online gateway for prospective students to apply. The system receives nearly 4 million applications per year. To save money and ease the management issues dealing with their original hosting provider, Sacramento based Infiniti Consulting, an AWS Education Competency Partner, migrated CCCApply to AWS. With a “lift and shift” from their previous cloud provider to AWS, Infiniti had CCCTC up and running and accepting applications in less than a day.

With the successful transition, CCCTC has experienced increased visibility into their data, specifically the amount of fake applications sent in daily. Unfortunately, the system is a target due to hackers wanting an “.edu” email account to claim a $1,500 federal education tax credit. They identified that more than 10,000 applications per month are fake, costing the colleges valuable time and money sifting through applications. CCCTC decided to utilize Infiniti’s machine learning and predictive analytics competency to help tackle the fraud risk.

To start the analytics process, the application was moved from the operational data store into simple Amazon Simple Storage Service (Amazon S3) buckets to create a data lake. From there, Infiniti data scientists applied their machine learning model and predictive analytics experience, in conjunction with AWS, to analyze what exactly was going on.

Infiniti’s use of AWS scalable infrastructure and machine learning tools means submitted applications are now scanned and characterized as either “good” or “suspect” before the college even knows they received an application. Suspect applications are quarantined and are available to the colleges for review should they wish. The goal is to save thousands of hours per year of valuable college admissions department staff time. Infiniti is also looking to build and implement a ML SPAM filter that can do a better job of identifying fraudulent applications and train itself to keep pace with fraud as it evolves and adjusts to the improved SPAM filter.

“Infiniti and AWS allow us to take CCCApply 288 data elements and turn them into something that is query-able, not just an excel spreadsheet,” said Lou. “We can now make decisions based on actual data to improve both the employee and the student experience.”

In addition to tackling fraud, Lou and his team are also diving into the admissions data to help increase student retention. They want to make consistent recommendations to students that can improve their experience at CCC. From on-campus childcare to study groups, machine learning helps connect the dots for students and make their time at CCC more valuable.

“Infiniti is proud to be CCCTC’s cloud services partner in delivering cloud applications to the California Community College system and the students they educate. We are committed to combining leading edge cloud expertise with AWS highly scalable and efficient cloud infrastructure to help support the community college system,” said Infiniti’s President, Scott Drossos.