Alumnus Using His Linguistics Degree as a Language Engineer at Amazon

Patrick Kelley received his Ph.D. in Linguistics from MSU in May 2018 and currently works at Amazon as a language engineer, where he works with technologies to improve Amazon’s Alexa, a form of artificial intelligence.

Could you give a basic job description of your position at Amazon?

I am a language engineer at Amazon, where my day-to-day changes almost every day, which I really enjoy. I work with technologies to improve and innovate Alexa, and my skills as a linguist are incredibly useful for working through complex language problems. Some days, I am in meetings all day long, and others I am planted in front of a computer writing code and documentation. I also have opportunities to travel the world to talk with other scientists about artificial intelligence and machine learning.

What did you originally plan to do with your linguistics degree? Did you ever imagine you would be working at a company like Amazon?

While I was a graduate student at MSU, I didn't ever imagine I would be working at a company like Amazon. Linguists, or at least as I was told from peers, typically went in a select few paths: pursue academia further with post docs and professorships, government or legal work, and industry, with this category being particularly constrained to translation. However, thanks to my guidance from my advising professor, Dr. Alan Beretta, I was gaining a ton of experience working with computers and operating a neurolinguistics lab, which set me up perfectly to pursue a career in machine learning and computational linguistics. As I dug deeper and deeper into artificial intelligence and speech recognition systems, the path before me was clear.

How did your time at MSU prepare you for your career?

MSU provided me an opportunity to lead and problem solve on my own. When I arrived at the neurolinguistics lab, we were operating at a very basic level. With some guidance from older graduate students, I was able to learn the core processes behind running neurolinguistic and psycholinguistic experiments, but the process was arduous and slow. As I got older and more experienced, I started to innovate the lab, bringing in new techniques, technologies, and code I wrote to make each step of our workflow more efficient and field standard for the rest of the team. With my advisors encouraging this self-driven exploration and innovation, it gave me the skillset to succeed at solving problems on my own and importantly the computational background I needed to succeed at a company like Amazon. With respect to my degree, coursework largely focused on theoretical linguistics, and while this was helpful in making me a well-rounded linguist, the experiences in my lab were truly where my skillset was honed. Importantly, my lab mates, advisors, and peers at graduate school were instrumental in helping me achieve my goals and making graduate life so enjoyable.

What was your research focus when you were working on your Ph.D. at MSU?

I specialized in Neurolinguistics, where I wrote a dissertation on the processing of a class of sentences called grammatical illusions, specifically Escher Sentences. These are sentences like, "more people have been to Russia than I have," where the sentence has no meaning (X people have been to Russia vs. I have been to Russia, True or False), and yet many find these sentences okay. I tried to figure out why.

When was it that you decided to study linguistics?

When I started my freshman year of undergrad at the University of Michigan, I was not really sure what I was going to do. At orientation, I was originally slated to take biochemistry, organic chemistry, calculus, and biology. At the time, I realized that I didn't really want to take any of those courses, so I scrapped them all and instead enrolled in physics of music, creative writing, anthropology, and brain and language. The brain and language class particularly piqued my interest, as I had never even heard of linguistics before. Once I started to be the annoying student who asked a million questions in class, the professor caught on and encouraged me to sign up for a major, so I declared my first semester of college.

At that point, I didn't really have any good idea of what to do with my linguistics degree, and as I approached my senior year, I was told that I really needed to pursue graduate school. At first, I wasn't so sure, since five more years of school didn't sound appealing at the time. However, once I started, I knew I made the right decision, and I loved every minute of it.

What is the most rewarding part of studying linguistics?

To me, the most rewarding part is that we study something that many people take for granted. Oftentimes, language is considered something you just use to communicate, and we typically don't think too deeply about language, since most people can perform language without thinking too deeply about it. But, at the basic level, what we are doing almost every day is an incredible feat: we are able to, in an already noisy environment, take a bunch of weird sound waves coming from people's lungs and mouths, translate those physical acoustic disturbances into some meaning we conjure up in our brain, then we make a series of odd articulations with our tongue, mouth, throat, and lungs to produce a series of sounds that others can also understand, all within a blink of an eye. How are we able to do this so quickly and so accurately? How is it that babies and toddlers can figure this complex process out, without explicitly being taught how to, before they can do something simple like feed themselves or tie their shoes? As we start to explore the mental faculties that we take for granted as humans, we start to discover amazing feats and wonders within each and every person. Now that we are trying to emulate this process with machines, it has become clear: solving “language” is a hard problem, and it's an incredibly enriching experience to be working on trying to solve the many mysteries surrounding language.

What is something students may not be aware of surrounding the linguistics major?

Linguistics is a science. We utilize the same scientific principles that physicists, biologists, and chemists, etc., all use to gain more knowledge about how the human mind works. We don't just sit around and talk about how old languages are neat (though, to be fair, they are), and we don't necessarily speak 10 languages. I would encourage everyone to take a linguistics course at least once in their lifetime because I believe the knowledge we have to share will truly amaze and surprise you.

What types of opportunities does the linguistics degree offer?

Typically, linguistics is considered to be a very narrow field with not a lot of applications for jobs. But, as the world changes, so does the demand for different jobs. Now, machine learning, artificial intelligence, and speech sciences are highly sought after. So, if I were to advise new linguistics undergraduates or graduate students, I would recommend they start to work on their computer science skills since this route offers the broadest opportunities for a career after school (plus, it is super exciting and interesting). The harsh truth is that if you would like to pursue academia, there are not many jobs available, and a post-doc is almost always required. Other opportunities a linguist could pursue would be education, translation, intelligence work, speech pathology, or if they branch into the cognitive sciences, other kinds of research opportunities. This, of course, is not a limited list, since linguistics itself is a holistic discipline, I have seen fellow linguists branch off into very diverse fields. I strongly believe that the assumption of the jobless linguist is fading away.