Jelani Nelson: A self-starter and a pioneer in the field of algorithms

Thank you very much for agreeing to do this interview with us. Can you please introduce yourself to our readers? What do you do?

I’m a computer science professor at UC Berkeley, with a research focus on algorithms (especially for large and/or high-dimensional datasets). My broad home area is theoretical computer science. I also recently started working at Google one day per week as a research scientist to put some of my theory knowledge into practical use.

Can you tell us about your education background?

I went to a Montessori School in Phillips Ranch (Southern California) for pre-k and kindergarten. I don’t have too many memories from that school since I was young, but I know they gave me a tremendous headstart: I came out of kindergarten reading novels and knowing my multiplication tables. I then went to All Saints Cathedral School in St. Thomas, US Virgin Islands from 1^st to 12^th grade, which was also a fantastic school, and I credit my teachers and school environment there for much of my success. I then went to MIT for undergrad (math and computer science double major), and then a Master’s and PhD in computer science. That was also a great place filled with incredible people and many opportunities to learn and grow.

How challenging was it to get admitted to one of the finest engineering schools in the world?

That’s a hard question for me to answer. Senior year of high school my dad brought home a US News college ranking magazine; at the time (Fall 2000) I saw that MIT was ranked 1^st separately in both math and computer science. I had never heard of the school and also didn’t fully appreciate how hard it was to get in, so I only applied to MIT. That was a very risky thing to do (knowing what I know now, I don’t recommend it), but luckily it worked out and I got in. Only years later when talking to others did I realize that applying to only one school is not recommended, and I then appreciated how lucky I had been.

Did you have to pay for your studies? Did you have any financial burden on your education journey?

I was very fortunate to grow up in a household where money wasn’t ever really a problem. My parents made too much for me or my siblings to qualify for any financial aid; they just paid undergrad tuition cash for all of us (and medical school for two of my sisters who went that route). My Master’s year and PhD were fully paid via research and teaching assistantships and fellowships, and I even received a stipend that allowed me to be financially independent from my Master’s year onward. I also did seven summer internships throughout undergrad and grad school, which provided a nice bit of extra income that I invested.

Looking back, it’s really amazing to me how big a difference that gift from my parents made in my life, i.e. not having any student loan debt. It allowed me to buy a house with my own savings immediately when I started my first job (as a professor at Harvard), which I eventually sold to profit more than double what I put into it. When I eventually married my wife, it turned out she did her medical school in Ethiopia where tertiary education is 100% free, so she didn’t have to worry about debt either. It’s been a huge blessing. If America is serious about equity and leveling the playing field, we need to drastically reduce the cost of tertiary education in this country.

How did you find your passion for Computer Science?

I’d say it started with video games. I first played a lot of console games, then switched to computer games. Playing computer games got me into using the Internet, which led me to being curious about HTML and teaching it to myself in middle school, then teaching myself some C/C++ later in high school. I decide before entering college, i.e. some time in the late ‘90s, that I was going to double major in Math and CS.

There are people who believe in the “geek gene” hypothesis claiming some people have the 'natural talent' to excel in computing sciences. What would you say to those people?

Computer science is broad, and there might be some pockets where it helps to have some natural talent, especially in the more mathematical parts of the field. But I think generally, most people can have a successful computer science career with or without that gene, if it exists.

What is your research about, and what kind of problems does your work solve?

I work a bit on sketching algorithms, which are algorithms that summarize very large datasets using very little memory while still capturing useful information for later queries. We usually want these summaries to be aggregable across different nodes in distributed computing environments; for example, imagine I’m a large Internet company with a fleet of servers serving content and ads to users. I then might want to be able to answer queries like “how many distinct users saw ad X across ads served by my entire fleet of servers combined?”. By exchanging only these short summaries, or sketches, the servers can minimize communication bandwidth. This is just one example, but sketches are used in many applications not only to reduce communication, but also to for speed and memory usage improvements. I also often work on dimensionality reduction problems, where we have some high-dimensional data (e.g. in machine learning, where they talk about having many “features” of the data), and we want to reduce dimensionality in a pre-processing step to for example improve runtimes.

I know that you have received multiple awards, such as the Presidential Early Career Award and the George M. Sprowls Award for Outstanding Doctoral Thesis. What motivates you to excel and break barriers?

People ask me how I type so fast (I was once ranked the fastest typist in the world on typeracer.com and typera.net). Part of the answer is that I obsess with self-improvement – I guess you could say I’m a perfectionist. There were phases in my life 20 years ago when I just typed texts over and over and over again for hours on typing websites, every day, to gradually get better and better so that I could keep breaking my own previous records. That got me from 120wpm as a 12-year old to eventually hitting 232wpm in my early 20s. It was a slow and gradual process, but there was consistent progress.

So to answer your question: I think that ability to obsess has helped me in many other ways as well. When I was 18 I joined a Japanese gaming clan online and bought a Japanese textbook to self-study – I probably devoted 6+ hours to self-teaching Japanese every single day that summer. That allowed me to skip a year of Japanese at MIT. In computer science, I did the same thing the next summer but self-training algorithms, solving hundreds of problems one summer in my spare time. Research during Ph.D. was similar. I obsessed about research problems.

What was the hardest hurdle you encountered in your career? How did you overcome it?

The only hurdle I can really identify was Freshman year of MIT: I realized I was at a disadvantage compared to some of the top Freshmen who had done math and computer science olympiads in high school and simply knew much more than I did. I overcame it by talking a lot to those very strong students and trying to learn from them.

Where did you grow up and how was it like growing up there?

St. Thomas, US Virgin Islands. Well, it’s paradise of course. We have the best beaches in the world, and my community was very supportive of the youth in education. We had some of our own problems, but we didn’t have many of the major problems that plague the mainland USA, e.g. race-related issues.

You have founded AddisCoder, a free summer program in Addis Ababa, Ethiopia. Can you please tell us more about the program and what inspired you to start it?

I started it the summer after finishing my Ph.D. I was planning to visit Addis Ababa anyway that summer to see some friends and relatives, and I figured I’d do something on the side to keep myself occupied while there. I was loosely inspired by the MIT Africa IT Initiative (“AITI”), which had been started years earlier by three African international students at MIT (one Ethiopian and two Kenyans). Their program taught computer science and entrepreneurship to undergraduates at various African universities, including at Addis Ababa University AaIT (“5 Kilo”). I first thought about teaching a special topics course at the university that summer, but it turned out to be logistically easier for me to run a high school program so I did that instead.

How was your experience teaching students in the AddisCoder program? What were your expectations? To what extent were your expectations met?

I was a bit too ambitious. All the students but one had never programmed a day in their lives, yet my goal was to teach them roughly 2-3 MIT computer science courses in 4 weeks. I figured we were meeting five days a week and they had no other courses taking up their time, so it was worth a shot. I probably covered more than I should have, so students weren’t able to absorb everything, but I do think most students got a lot out the experience and learned a lot. They were really smart, and many of them ended up pursuing computer science after the program and doing very well.

As someone who has taught students both in top US universities and in Ethiopia, how related were those experiences? Any interesting observations?

It’s hard to compare. In the US I teach undergraduate and graduate students. In Ethiopia I’ve only taught high schoolers, the vast majority of whom had never programmed a line of code before my course. I would say the hunger to learn is universal though across all the places I’ve been. I would say there’s a bit of a difference in terms of access to opportunities though. The top high schoolers in the USA are exposed to college-level material and even research (e.g. RSI and the ISEF) at a very young age, and some of them enter freshman year of college already looking for more research opportunities. That’s highly uncommon in Ethiopia.

People of African descent are underrepresented in STEM fields. How can we motivate more people to get into STEM?

I think the first step is to not ask “what can we do”, but rather “what can I do”. (Who is this nebulous “we” anyway?) And there are many fine answers to that: invest in businesses, mentor someone, teach a course, give motivational talks, etc.

AddisCoder has done an astonishing job in supporting its students to join some of the best schools in the world. What do you think is the main driver of the success of the program?

Ethiopia is a big country, over 110 million population, and we recruit some of the best students in the country. That helps a lot. We started small, just recruiting students from Addis Ababa, but starting in 2016 we began co-organizing the program with the Meles Zenawi Foundation (MZF), and they helped a lot in expanding our recruitment to the entire country. We also have been fortunate to recruit some amazing teaching assistants over the years, all of whom were strong computer scientists, and many of whom either had or were currently pursuing PhDs, had competitive programming experience, or had industry experience at top software companies like Google, Microsoft, Facebook, and Dropbox (just to name a few).

How can excellence in STEM help Africa overcome its challenges?

Less than 10 years ago, when the Swedish Royal Academy announced the Nobel Prize winners in Chemistry, they stated “Today the computer is just as important a tool for chemists as the test tube”. We have companies like Gro Intelligence using computation and big data for applications in agriculture. The US Census Bureau now uses differential privacy, a notion invented by theoretical computer scientists, to protect the privacy of citizens in census data. The economy not just of the future, but of right now, is built on computation, and those economies which invest in computer science and STEM more generally will be the winners.

What are important lessons you've learned on your way to where you are now in your career?

Make sure to maintain good friendships and enjoy your life.

What is something that someone wouldn't know by looking at your profile? Any fun facts?

Once when one of my AddisCoder students was denied a visa to attend a summer program in London, me and a colleague tried to reach the Queen of England to get her to invite him as her guest so he’d get the visa. We didn’t reach her, but we actually got decently far. It’s really too bad for the UK; they missed out big time. He was one of the strongest kids in our program.

Finally, what are your pieces of advice for the African youth?

The same advice I give all youth: self-educate. ccw.mit.edu, coursera.com, and edx.org are all great resources.