Posting Description

POSTDOCTORAL ASSOCIATE, DORKENWALD LAB, McGovern Institute for Brain Research -  Dorkenwald laboratory, to develop computational reconstruction, analysis, and modeling approaches for connectomics datasets. The lab develops computational approaches to reconstruct, analyze, and model large-scale connectomes, aiming to uncover organizational principles of neuronal circuits and how circuit structure supports computation. Will lead research on one or more of the following areas: Automated proofreading & annotation at scale: Machine learning approaches for error detection, human-in-the-loop proofreading of automated cell reconstructions, active-learning approaches for efficient annotation, and self-supervision approaches for tokenizing image datasets and cell reconstructions; Circuit analysis & modeling: Analysis of cortical connectomes, including comparative analyses across ages/regions; hypothesis-driven tests of discovered circuit rules; pair analyses with data-constrained models (e.g., RNNs, dynamical systems) and simulations; Morphology representation & multi-modal linking: Learn representations of detailed cell morphologies to link across datasets (within connectomics) and across modalities (e.g., EM ↔ Patch-seq) to build multi-modal connectomic resources that provide the basis for analyses that combine, e.g., connectivity with transcriptomic information; and publish in leading venues, maintain high-quality, reproducible code, collaborate across McGovern/BCS and external collaborators, and mentor students.