Jelani Nelson: The Ethiopian-American Scientist Behind the Algorithms That Power Big Tech

Every time Google counts how many unique visitors landed on a page, or Facebook tracks distinct IP addresses in a data stream, there's a good chance an algorithm Prof. Jelani helped build is doing the work. The UC Berkeley professor's research doesn't make headlines the way AI chatbots do, but it sits quietly inside the infrastructure that keeps large-scale digital systems running.

Prof. Jelani was born in Los Angeles to an Ethiopian mother and an African-American father, then grew up in St. Thomas in the U.S. Virgin Islands. He taught himself to code the way a lot of curious kids did in the early internet era: right-clicking on a webpage, seeing "view source," and going from there. By 11, he was writing HTML. By high school, he had moved on to C and C++.

He went to MIT for his undergraduate degree, double-majoring in mathematics and computer science, then stayed for his PhD. His doctoral work focused on streaming algorithms, a field concerned with a deceptively simple problem: how do you answer questions about enormous datasets without storing all of it?

The Science of Shrinking Data

Companies like Google and Meta have petabytes of data flowing into their servers every minute. Storing and querying all of it in real time would be computationally expensive to the point of being impractical. Prof. Jelani's work on sketching algorithms addresses this directly. A sketch is a compressed summary of a dataset, often exponentially smaller than the original, that still lets you answer specific questions accurately.

One of his best-known contributions is helping prove that the Johnson-Lindenstrauss lemma is optimal, alongside Kasper Green Larsen. He also co-developed the first theoretically optimal low-memory algorithm for counting distinct elements in a data stream, a problem that sounds trivial until you're doing it at internet scale.

For his research, he received the Presidential Early Career Award for Scientists and Engineers, an Alfred P. Sloan Fellowship, and an Office of Naval Research Young Investigator Award.

AddisCoder: A Side Project That Became Something Bigger

In the summer of 2011, Prof. Jelani finished his PhD and had a gap before his postdoc at Berkeley started. He decided to visit relatives in Addis Ababa. He figured he'd teach something while he was there.

What started as one lecturer, one volunteer teaching assistant, and 82 students in a borrowed university classroom has grown into AddisCoder, a nonprofit that has now trained close to 700 Ethiopian high school students in algorithms and programming. The program is free and residential, covering food, housing, and transport so that students from all eleven of Ethiopia's regions can attend regardless of their family's income.

Over 40% of participants are girls. Alumni have gone on to study at Harvard, MIT, Stanford, Princeton, and Columbia, and some have joined Google. In 2023, the Association for Computing Machinery gave Prof. Jelani the Eugene L. Lawler Award for humanitarian contributions to computer science, citing AddisCoder by name.

In 2022, he co-launched JamCoders in Kingston, Jamaica, using the same model.

Why This Matters for Ethiopia

Ethiopia has a shortage of computer science educators who can teach advanced material. When Prof. Jelani first arrived in 2011, he had wanted to run a graduate-level course. That didn't come together, so he redirected to high schoolers. In hindsight, that pivot did more.

AddisCoder doesn't just teach coding. It works with Ethiopia's Ministry of Education to identify students from public schools who might otherwise never have access to this kind of training. The program has become a pipeline, connecting young Ethiopians to top universities and careers in computing that would have been out of reach.

Prof. Jelani's research and his education work look like separate tracks, but they're driven by the same instinct: find the most efficient solution to a hard problem, and make sure the people who need it can actually access it.

His algorithms reduce waste at a technical level. AddisCoder reduces a different kind of waste, the kind that happens when talented students never get a door opened for them. Both are problems worth solving.