The group that created Lobster AI was announced at CCI's Senior Showcase.
During the school year, a group of students at Drexel University at Computing & Informatics College (CCI) developed a new AI (AI) tool that helps streamline the world of medical books. To achieve this, they worked with industry partners AOIC, LLC, and medical and scientific communications companies to go further than expected in their senior projects.
“It was an ambitious project that students put in great effort,” said Mark Altieri, a financial and technology stakeholder who gave instructions to one group of students on technology requirements. “What impressed me most after technical expertise was that there was a really dynamic group that each brought expertise to the table. I never saw an unprepared student on the phone.
CCI seniors develop software projects through a year-long class, usually working with real-world professional partners to develop software projects. This means everything from budgeting and planning to planning, design and implementation. Sometimes the project is not in-depth to be fully completed in nine months, but it's as close to the real world project you can get, adding to the experience of students riding a cooperative. For students who have been at Drexel for five years, they get three six-month cooperatives working in their chosen field, still learning.
“If you add an 18-month cooperative to a nine-month senior project, you'll never know any other universities that have graduated from school with 27 months of experiential learning and do so,” said Jeff Salvage, a professor of computer science.
The student group included Leo Lee, Software Engineering '25. Abby Tabas, Computer Science '25; Maame Adwoa Ocran, Computer Science '25; Jason Morais, Software Engineering '25; and Viven Ho, Software Engineering '25. They were tasked with using AI to help develop the first draft of medical manuscripts. This is a formal document detailing the results of a medical examination or study. These include method sections that explain how the research was conducted, including research design, data collection, and data analysis. This can take time for writers despite its structured form. Sarah Hunmasti, Vice President of Science Strategy and Services at AOIC, which leads the medical writing team, said her team wanted to find a way to streamline the process. This project focused on a third phase clinical trial to test the safety and efficacy of new treatments compared to previous treatments. The trial has a protocol that lists everything about how to complete the study, but it can be hundreds of pages.
“If you want to share it with other scientific and medical communities for your manuscript, you need to condense it into something that provides all the important and important information,” says Hummasti.
The project was complicated by the data security and confidentiality protocol called AOIC Stewards, which meant that students were already working with data in the public domain. According to Altieri, it was real-world data, but from a manuscript that had already been written.
Hummasti provided examples of method sections and checklists that can be used to create checklists, with students training AI based on what the writers do. Building AI is a labor-intensive process due to confidentiality concerns, and the manuscript itself is usually very routine, but each test is different.
“It gave them a starting point and an endpoint so they could use it to train AI,” Hummasti said. “There are always some differences, but what is consistent, and how do we get home to them?
Students made progress each week, updating AOIC representatives at meetings, discussing their approaches and asking questions. Salvage prioritizes attracting students at stakeholders and fasciti time as it helps to develop soft skills essential to client relationships, such as communication, teamwork, professionalism, and time management. Additionally, they understand what a real-world client experience is, and their real-time real-world connections help students see what is happening in their industry at the moment.
“I think the atmosphere and the expectations Jeff has introduced are very important,” said Tom Burke, president and CEO of AOIC (and a two-time Drexel alumnus). “The class is clearly a serious, real-world focused class. Student participation and dedication are important and we will not achieve that unless the backbone of the program is built through faculty and university.”
Ultimately, students develop a tool known as Lobster AI, and throughout the year the group is far ahead of AOIC's expectations. As far as the AOIC team knows, this is a new product not developed to be sold by other companies, giving students a lot of experience in what innovations their employers are looking for when they head to the workplace.
“It's a novel approach and could be fruitful,” said Altieri, vice president of finance.
Currently, the Li, Tabas, Ho, Morais and Ocran groups have graduated and are now in the real world. As for AOIC, it coincides with another senior group that can take the tool to the next level from 2025 to 26. They know they can rely on Drexel students.
