Registration (Harvard Only, please use your Harvard Email to register)
Chester Curme is a data scientist at Kensho Technologies, where he serves as the technical lead for the Data Extraction team. Prior to joining Kensho, Chester worked as a quantitative analyst in fixed income at Loomis, Sayles & Co., an asset management firm in Boston. He holds a B.A. in Physics and Mathematics from Middlebury College and a Ph.D. in Physics from Boston University.
Abstract: Much of the world's data is unstructured. It lives in audio, images, and text arranged in free-form documents. Working with such data efficiently often poses technical challenges. This is particularly relevant for the investing world, where access to timely, clean, and accurate data can be crucial to a successful operation. Kensho Technologies, headquartered in Harvard Square and a subsidiary of S&P Global, develops software to extract structured information from unstructured sources. I will explore the host of technical challenges involved in building structured datasets for investors, and show how machine learning technologies can be harnessed to execute these tasks at unprecedented efficiencies.