AI Neuroscientist

Role overview

Set the scientific standards used to evaluate whether AI agents can reason through neuroscience research and make defensible research and development decisions. You will work with neuroscientists and software engineers to create realistic benchmark tasks spanning basic nervous-system function and disease, from interpreting evidence to choosing experiments and assessing conclusions.

Responsibilities

- Define expert answers and explain the evidence and reasoning behind them for neuroscience research questions and development decisions. - Design realistic scenarios covering hypothesis formation, target and model selection, experimental design, data processing, and inference. - Assemble evidence packages combining modalities such as electrophysiology, imaging, brain scans, omics, pathology, behavior, and biofluid markers. - Compare healthy function with disease-related changes to assess whether agents can transfer knowledge across contexts. - Assess agents’ choices of model systems and analysis units, including whether findings from laboratory models apply to human disease. - Examine reasoning step by step, identify unsupported conclusions or missed confounds, and judge how well confidence reflects the available evidence.

Requirements

- PhD or MD/PhD in neuroscience or a closely related field, with hands-on research, first-hand data analysis, and publications. - Direct experience analyzing raw data in at least two neuroscience modalities, such as electrophysiology, optical imaging, EEG, neuroimaging, omics, neuropathology, or quantitative behavior. - Deep expertise in at least one neuroscience area and working fluency across biological levels, from molecular and cellular processes to circuits and behavior. - Strong command of experimental design and statistics, plus practical understanding of model systems including cell-based models, organoids, animals, and human cohorts. - Experience using AI agents and the scientific judgment to catch reasoning that sounds technically plausible but is not supported by evidence. - Ability to challenge an agent’s conclusion, communicate uncertainty, and revise a judgment when the evidence warrants it.

Nice to have

- Familiarity with neuroscience and biomedical data resources, such as brain cell atlases, standardized electrophysiology and imaging datasets, and major disease or population cohorts.

Benefits and work setup

- Remote work is available; onsite work is preferred at an office in San Francisco. - The listing describes performance-linked compensation and different employment arrangements and benefits for US-based employees and international contractors. It also outlines individual contributor and management growth paths.