Data Science Capstone Connects Students with Industry
By Kirsten HeuringMedia Inquiries
- Associate Dean for Communications, MCS
The inaugural class of the Mellon College of Science Master of Science in Data Analytics for Science (MS-DAS) program graduated this May. As part of the program, students participate in a capstone project, where they work with partner companies to get hands-on experience in data science.
"Unlike many MS programs in data analytics, the MS-DAS program at MCS is quite unique because it is tailored for students with backgrounds in the foundational sciences, such as biology, chemistry, math or physics, to acquire additional skills in many aspects of data science and gain real-world experience through the capstone project course with industry partners,” said Manfred Paulini, associate dean for faculty and graduate affairs and professor of physics.
Bristol Myers Squibb (BMS), a global pharmaceutical company that develops innovative medicines to help patients prevail over serious diseases, provided a valuable opportunity for the MS-DAS students through a Drug Product Development (DPD) sponsored initiative led by Joseph Valente. Such opportunities are part of a broader effort within DPD-BMS to establish a strong rapport with key universities and their students and become a destination of choice for top talent. Marilia Barros, a physics Ph.D. alumna who is now a Principal Scientist at BMS, reached out to Carnegie Mellon University at just the right time.
“Recently, the application of data science and data analytics has become increasingly important for rationalizing experimental datasets and unlocking trapped knowledge,” said Barros. “Beyond impacting learning and professional development of the next generation of scientists, BMS would benefit from the application of modern data analytics techniques to industry data sets in alignment with DPD digitalization initiative and strategic priorities. Being able to partner with CMU and have their students working on a scientific project brings a fresh perspective on current questions and problems relevant to BMS.”
Barros and two other BMS scientists, Stephen Thomas and Andrew Ilott, provided a team of two students, Edward Piechowicz and Jonathan Taylor, with a data science problem and mentored them as the capstone progressed.
Piechowicz and Taylor were given the following problem statement: leverage in silico reaction modeling approaches to supplement experimental excipient (non-active compounds that enhance product stability and processing) compatibility studies that are needed to support formulation development of small molecule active pharmaceutical ingredients. Based on this information, the team was able to create a model to predict chemical structures of degradation products, resulting in actionable insights for formulation development.
“I think it's been a phenomenal experience,” said Piechowicz. “Our partners in the company have been incredibly helpful and supportive for anything we might need whether that be technical advice or domain knowledge. It's been a really great experience to have the opportunity to take what we've been learning in our classes and apply it to a real-world example and a real-world problem.”
Besides learning more of the technical aspects of data science, Piechowicz and Taylor were also able to develop other skills.
“I think probably the most valuable thing was just learning how to communicate science outside of academia,” said Taylor. “That's something that we've been putting a lot of emphasis on. I've learned to really communicate science and communicate results to other professional scientists.”
Both Taylor and Piechowicz secured full-time positions prior to graduating. Taylor is working as a data analyst at a small biotech company in San Diego, and Piechowicz accepted a position as a scientist on the data management team of an environmental consulting firm in Albuquerque.