Fabio Urbina

Throughout Fabio's career, he has combined computational tools and machine learning with classical small-molecule, molecular, and cell biology techniques to address previously-difficult to probe scientific problems. He excels at finding solutions to drug discovery with machine learning, as well as devising and validating semi-automated or automated methods to improve the throughput-capabilities of currently implemented experimental pipelines.

During his graduate student career, he developed a computation image analysis tool which automatically identified exocytic events from neuronal imaging, a task often performed by hand, turning a subjective and time-consuming task into a high-throughput, objective process. Over the course of his work, he became passionate about complex data mining tasks and new research in image analysis.

At Collaborations Pharmaceuticals, Inc., he has applied his skills to semi-automate extractions of complete datasets for use with machine learning from large databases of experimental results, a task previously found to be exceedingly difficult. Using his up-to-date knowledge on machine learning techniques and experienced coding ability, he rapidly adapted and test new machine learning models (<1 year old publication dates), allowing him to evaluate new frameworks as they are introduced in literature. He has solved complex data workflows, automating > 1-terrabyte datasets through machine learning pipelines with reasonable turnaround time. I’ve added new software additions to the Collaborations Pharmaceuticals, Inc. library, including new generative models for de novo molecular design and a refactoring our existing Assay Central software to be microservices oriented. As Associate Director, he lead a team of machine learning researchers and software engineers to build a CI/CD pipeline and develop novel LLM and deep-learning frameworks for predicting new drugs which we have validated in the lab.

With a skill set of molecular, computational, and theoretical machine learning techniques, he has unique insight into limitations and strengths of where machine learning can lead to value and can bridge the gap between subject matter experts and machine learning researchers. His knowledge of the current machine learning literature and ability to swiftly adopt and deploy new technology makes me an ideal candidate for the integration, testing, and deployment of Machine-learning based pipelines.