Jelani Nelson, an assistant professor of computer science, joined SEAS in 2013. His research focuses on the development of efficient algorithms for massive datasets, especially algorithms that use little memory and require only a single or few passes over the data (so-called “streaming algorithms”). 

Using a technique called "sketching," it is possible to create a very compressed version of a dataset that lacks all the detail but provides sufficient information to answer certain queries. Sometimes, these sketches can be exponentially smaller than the original dataset, allowing savings in processing time, communications bandwidth, and required storage capacity. The algorithms that produce these sketches, however, must be carefully designed to guarantee an adequate degree of accuracy.