Developing AI to Capitalize on IoT Systems

Amisha Agarwal (MSIT '21) was a first-quarter student in Northwestern Engineering’s Master of Science in Information Technology (MSIT) program in 2020 when she learned of a potential internship at Northwestern’s Center for Deep Learning (CDL). The selected intern would work with Professor Diego Klabjan, the director of the CDL, to develop an end-to-end software stack to capitalize on Internet of Things (IOT) systems, state-of-the-art artificial intelligence, and machine learning to produce predictive results.

Agarwal thought the internship sounded amazing, so she applied. Her background was in software and web development, but she was told she did not have enough of the data science experience required for the position. Undeterred, Agarwal shaped her MSIT experience to include courses and projects that helped her develop a good understanding of basic concepts surrounding machine learning and analytics, including MSIT 431 - Introduction to Statistics & Data Analysis and MSIT 423 - Data Science for Business Intelligence. 

In March 2021, Agarwal reached out to Professor Klabjan again with her updated resume and  strong desire to work on the project. 

“Seeing my eagerness to work, he gave me an opportunity to be a part of the project for a month to analyze and judge my capabilities,” she said. “After being impressed with my performance during this period, he offered me a full-time internship.” 

Agarwal said the project, called REFIT, is challenging, but that's what makes it interesting. REFIT is built on a collection of open source platforms, most of which Agarwal never used before. The goal of REFIT is to make it possible to quickly develop, experiment, and deploy a robust, powerful, feature-rich IoT framework that makes work more efficient without adding any vendor lock-in for the end users. 

Agarwal's programming experience helped speed up her understanding of the different platforms she was expected to use. Reading about the software helped, but the best way for her to learn was to be hands-on with the platforms. 

"The most efficient way to learn new technology on the job, apart from reading the basic literature, is to debug and solve more errors and issues," she said.  

Agarwal's technical knowledge grew immensely during the internship, but she learned more than just about new platforms. During her time with REFIT, she was assigned additional responsibilities, including leading the team of other interns, giving her first-hand leadership training and practice. She had the opportunity to interact with potential partners and subscribers to the product, providing her with real-world examples of how to connect and interact with current and prospective clients. She also was selected to represent the work done on the project at October's IEEE International Conference, where she will present on "Artificial Intelligence for Internet of Things." 

That varied experience is what initially made MSIT an appealing option for Agarwal. Beyond the internship opportunity, she was excited that MSIT’s course structure has a 70-30 split between technology and business courses and that students have the option to take courses across the university. 

“The idea of studying alongside part-time students was another important factor,” she said. “This program helped me build my network among professionals and learn from their ongoing experiences.”

With her internship complete and her final quarter underway, Agarwal has her sights set on life after MSIT. She’s still exploring what field of data science she would like to pursue in her career, but she knows it will build upon her experience interning at CDL. 

“I want to work with someone who envisions developing game-changing artificial intelligence tools,” she said. “I want to create fruitful solutions and help important decision-makers in healthcare, finance, or manufacturing.”