Anj Raman ’24

Whooo’s Reading, San Diego, CA

I worked for Whooo’s Reading this summer. The company provides a product to school districts, teachers and students and a platform for teachers to assess if students are doing their reading (hence the name Whooo’s Reading). It does this by grading student quizzes, as well as a couple of other metrics. The quizzes consist of fill-in-the-blank questions and those that require an open-response paragraph. There are also several other features on the site, including question builders that have students write their own responses to the questions. Whooo’s Reading makes this engaging to students by gamification. They allow students to purchase items for their owl mascot, including hats, articles of clothing and houses, and this encourages students to keep taking quizzes and earning points.

The first half of my internship involved looking at question builders (where students write fill-in-the-blank questions about books) and determining if we could repurpose them to use as potential questions we could ask other students in the future. I used a machine model (BERT) to grade student responses and analyzed the data to understand the significance of scores and suggest actionable insights based on the scores. I tried to see if any other variables were good at predicting the score (length of question, assigned) and then moved onto part of speech to see if that impacted the score. I found that certain variables, such as grade, impacted the score significantly. However, other factors did not influence the score at all. This finding was helpful to the company to further develop its products and provide a technical differentiation to its customers by providing further insight into their student’s learning.

The second half of my internship focused on the generic fill-in-the-blank questions on quizzes, which are asked when the company doesn’t have any background on the book. I created a script that asks for the book and outputs the main concepts that are generated from student responses to generic fill-in-the-blank questions. I used natural language processing and learned how to automate processes I never would have thought of before. This script will be useful to the company later as they can now extract the main idea from books without even having to read them, as well as grade student responses more accurately.

I learned a lot technically, which provided me a deeper look into the company such as what it means to be a data scientist and how to convert my theoretical work to processes that would be useful to the business. I was able to challenge myself with algorithm and software development. I also loved the people at Whooo’s Reading and the structure of the internship. I hope to continue learning and would love to do something similar in the future.

I extend my greatest gratitude to the ’68 Center for Career Exploration and to Bill McCalpin ’79 for this opportunity this summer.