Forget the comfort zone. For Hannah Jiao, a second-year master’s student in the Luddy School of Informatics, Computing and Engineering, it’s time to learn a new way of thinking and solve real-world problems.
And, hopefully, find a great job.
Welcome to the client-based collaborative project between Salesforce and the Luddy School’s AI Development and Experience Lab. With help from a pair of Salesforce employees and Salesforce LAIDEL faculty fellows Sai Shruthi Chivukula, visiting assistant professor, and Damir Cavar, associate professor of data science, the project guides students to develop new AI applications or apply existing AI technology to new problems and meet business needs.
Students work in teams led by coaches from Salesforce, who present them with current business problems to solve. Coaches provide guidance throughout the project. Chivukula and Cavar are mentors.
“This is a good opportunity so we can understand how a big company works,” says Jiao, who is studying human computer interaction.
Salesforce, Inc. provides customer relationship management software and applications focused on sales, customer service, marketing automation, analytics and application development.
LAIDEL helps students develop the knowledge and skill to do that.
“This semester is going incredibly well,” says Brandon Brackett, Salesforce director of digital experience strategy. “The students are asking all the right questions, and tackling each step in the curriculum just how we would in our day-to-day jobs.
“I hope that the students can see similarities in what they've learned through their academic journey and how they apply to real-world problems. The attention to detail and overall understanding of the project has been fantastic to see. I hope everyone is having as much fun as I am!"
There are two main projects -- the Trailhead Project and the Einstein Prediction Builder. Students are divided into five groups.
The Trailhead Project focuses on integrating two different websites (Trailhead and Salesforce). The central question -- how can the Trailhead site be easily accessed on the Salesforce site?
The Einstein Prediction Builder, which includes one data science student for each group, is designed to make integration between Trailhead and Salesforce easy to build for those without technical backgrounds.
“The students have the plan,” Chivukula says. “They have all the resources they need. They can ask Salesforce clients nitty-gritty questions.”
Adds Jiao: “We have a plan that is clear and helpful.”
Jiao is one of 18 second-year master’s students participating in the 16-week course, which involves three progress presentations and a final presentation.
Th goal is to provide educational experiences far beyond a classroom setting. The biggest benefit is that “students are getting client interaction while they’re in school.”
“Most have done internships, but many did not have business experience before,” Chivukula says. “This is a trial run for them to understand that this is how it works. There is a pace they have to understand.”
In other words, there are deadlines to meet and pressures to overcome.
“Now they’re getting the practical side of things,” Chivukula says. When that happens, she adds, “decision-making changes.”
“That changes the students’ perspective. They need to constantly see if the project is feasible. They don’t do that in a class setting. It builds their design thinking in a much more practical way.”
Beyond that, Chivukula says, it adds to their portfolios.
“That ups their games. All of them will apply for jobs. This is a project with a title with a company name rather than just a class project.”
Students are connected online with Salesforce coaches. Instruction is virtual only. Weekly plans are given, presentations are required, and collaboration tops the priority list.
“In industry,” Chivukula says, “you don’t work separately, you work together.”
The emphasis is to take what’s done in classroom settings, where students are free to try multiple approaches, and blend them into real-work situations where they target the most efficient and cost-effective approach.
“They get feedback to improve their technique and make this product better,” Chivukula. “In the classroom, there’s no real-world impact of what they create. It’s only learning focused.”
In business, impact is crucial.
“The product is real,” Chivukula says. “There are millions of people using it. There is accountability. Students have to restrain themselves and learn that balance.”
Cavir works with students to set up natural language processing and AI pipelines on products to map out information into a specific graph model. That will be used for machine learning.
He hopes to soon enter a phase of “real experimenting” with post-data preparation and algorithms setup, and concrete processing.
“We are developing interesting approaches to common problems known to Salesforce and other such companies,” Damir says. “The students seem well motivated and interested.”
Jiao says students want hands-on experience to learn how Salesforce employees work on product design. They want insight in not just how the product is designed, but why it’s designed the way it is.
“Students have to work with clients,” Chivukla says. “There are real business objectives, real money going on there.
“They also are working in groups and working remotely. Learning this skill is essential in today’s world. How do you work remotely with clients? How do you talk in company language? I wish every student would get this experience.”