Recent advancements in Environmental, Social, and Governance (ESG) considerations have increasingly influenced company shareholder meetings, where votes on these issues have become a critical mechanism for impact-oriented investors to exert positive influence through their voting choices. Consequently, a systematic examination of proxy voting records can yield significant insights into asset managers' ESG impact orientations. The project aspires to explore cutting-edge machine learning algorithms to accurately estimate the ESG positions of asset managers based on their observed voting behavior.