Next Gen Pharmacists Hunt for New Breakthrough Drugs with AI

At his desk at Genentech, Hanson Hoang, MS-AICD3, assesses machine learning models and looks for tools that will help scientists develop drugs more efficiently and accurately. The work feels familiar because it grew out of the capstone project he completed as a student in UCSF’s Master of Science program in Artificial Intelligence and Computational Drug Discovery and Development (AICD3).

“In my capstone project, I evaluated various machine learning methods for analyzing patient survival data across multiple cancer indications. Now, I’m working to develop AI tools and models, such as trying to reduce the manual effort required for routine tasks, such as generating documentation,” says Hoang. “Across both bodies of work, I’m always trying to learn more of the science while simultaneously helping scientists work more efficiently so they in turn can spend more time on the science itself.”

Hoang entered UCSF with a bachelor’s degree in biology with specialization in bioinformatics from UC San Diego and a strong computational background. What he didn’t yet understand was how those technical skills could be applied across the complex processes of discovering, developing, and testing new medicines, which is what he hoped to learn in the AICD3 program.

“Before AICD3, I didn’t know what the drug discovery or drug development process entailed,” he said. “The program broadened my horizons as I learned from those directly involved in the process and helped me realize this work is something I really enjoy being a part of and want to continue in the future. AICD3 gave me the foundational knowledge I needed to do the work I’m doing now at Genentech.”

His interest was well timed. Artificial intelligence is transforming the drug discovery pipeline, including identifying promising new drug targets — molecules in the body, typically proteins, that interact with medication to produce a therapeutic effect — speeding up clinical trials, and analyzing real-world data to make medicines safer and more effective after they’re approved. As a result, the biotech and pharmaceutical industries face a growing need for scientists who understand both AI and the biology behind developing medicines.

Training those scientists is the goal of AICD3, the School of Pharmacy’s five-quarter master’s program. It combines artificial intelligence, machine learning, data science, pharmacology, and drug development to train scientists capable of bridging two rapidly converging fields. The program is the first of its kind in the U.S. and welcomed its inaugural cohort in fall 2024.

“Interdisciplinary skills are going to be most important in drug discovery and development — we’re going to need to work at the intersection of biology and AI,” said Joanne Chun, PharmD, PhD, director of the AICD3 program. “Pharmacists already understand the science of medications. When you add AI and computational skills, that becomes an incredibly powerful combination.”

Rather than teaching AI in isolation, the program trains students to apply AI and computational methods across the full drug discovery and development pipeline — from target identification and molecular design through clinical development, regulatory science, and real-world evidence.

Learn and work with industry scientists

The field is evolving as quickly as AI itself, which is why the program is designed to evolve. Using project-based learning, students focus on real-world problems. “We need to be incredibly nimble because AI is changing so fast,” Chun said. “We’re constantly updating the curriculum to keep pace with the field.”

AICD3 students learn from scientists who apply AI every day at leading biotechnology and pharmaceutical companies. Industry experts regularly teach classes, present guest lectures, mentor students, and help shape the curriculum, while giving students opportunities to build professional relationships.

These accomplishments really demonstrate that this isn’t just a program teaching AI. It’s training a new generation of scientists who are already making meaningful contributions to the field.

Joanne Chun, PharmD, PhD, Director of the AICD3 Program

“Nearly every week we’re bringing in scientists from industry to teach the skills they’re using right now and share the challenges they’re working on,” said Chun. “Students have an opportunity to build relationships with industry scientists long before they sit down for a formal job interview.”

“You’re not just studying for an exam and then forgetting it. We’re building something meaningful that can actually be used,” said current student Tara Pande, MS-AICD3.

Students apply what they’ve learned to build AI models, analyze biological data, solve industry-inspired challenges, and complete capstone projects with industry or academic partners.

One of Pande’s most meaningful experiences was when she and a fellow student entered a UCSF AI health education hackathon with an idea inspired by attending lectures from other disciplines at UCSF. They proposed an AI-powered tool that could listen to highly technical scientific seminars and translate them based on a listener’s background knowledge. Their project earned second place and won the audience choice award, along with funding so they could continue developing the AI agent.

AI that hunts for cancer’s weakness

While both at Genentech, Huang and Pande are taking different approaches to applying AI to drug development. Huang’s work takes a systems-level approach to streamlining processes and Pande is focused on scientific discovery.

Pande’s capstone is working to understand why colorectal cancers become resistant to treatment and how to counter that resistance, using AI systems known as counterfactual generative models that simulate how biological changes might affect tumors. After spending some time working in the industry, she wants to pursue a PhD focused on computational models that simulate biological systems so scientists can work “in silico” — using a digital model of a cell to test how thousands of genes and proteins interact inside it. Her hope is that this will help researchers identify new therapeutic targets, especially ones that will be effective on specific patients.

“The AICD3 program as a whole, and classes like pharmacokinetics/pharmacodynamics more specifically, taught me what drugs do and why and how they do it. We learned about how AI can be applied to every stage of that process, and I think that really helped me figure out what I want to do,” said Pande.

Since the program launched in fall 2024, AICD3 students have won multiple AI and innovation hackathons, competed against participants from around the world in global AI and innovation competitions, published research, presented at national scientific meetings, completed capstone projects with industry and academic partners tackling real-world drug discovery and development challenges, and secured positions at leading pharmaceutical and biotechnology companies.

“These accomplishments really demonstrate that this isn’t just a program teaching AI,” says Chun. “It’s training a new generation of scientists who are already making meaningful contributions to the field.”