Characterizing biological foundation models

Inceptive is hiring a research scientist to characterize biological foundation models for AI-designed drugs. The role involves developing evaluations, analyzing model behavior, and collaborating with biologists. Requires PhD and experience in computational biology and ML.

• Embody our vision of an antedisciplinary environment and embrace learning about areas outside of your traditional area of expertise; Investigate how model performance changes with data quantity, data quality, dataset composition, and training methodology; Develop biologically meaningful evaluations and benchmarks that measure progress toward therapeutic design objectives; Design and execute rigorous experiments to understand the behavior, capabilities, and limitations of biological foundation models; Identify sources of potential artifacts, bias, and noise in biological datasets; Identify promising biological datasets for model training and evaluation, and develop computational pipelines for preprocessing, quality control, and exploratory analysis; Design studies, in silico or in the lab, that reveal what models have learned and which biological signals drive model behavior; Work with biologists to formulate hypotheses and translate biological questions into measurable machine learning experiments; Partner with AI researchers and engineers to prioritize research directions, data collection efforts, and model improvements; Analyze, visualize, and communicate experimental findings to inform decisions across teams