My interests nowadays is strongly situated in informatics. Yet, my path to this field has been rather untypical - or, as some people have said to me, "very untypical".

Earlier, I completed my bachelor's degree in Scandinavian Studies (2016) at Beijing Foreign Studies University, with a one-year state-funded exchange at the University of Southern Denmark. This is a study program with a mixed focus on language, literature and political science. As it is very unusual to study Scandinavian Studies in China (also, this study program only accepts around 15 new students every four years), I was often asked why I made this choice. Honestly, this was not a desired choice of mine at all: back to that time, I wished to study German linguistics due to my fascination with several rock/gothic bands from the German-speaking region. But in the higher education system in my country, especially contrasted with the German one where I currently reside, there is extremely limited freedom to choose the field of study or switch the study program. Instead, the priority was to play safe and secure a study place at any of the prestigious universities - no matter in which subject - rather than to choose a subject that follows one's personal interests. Due to several complicated reasons related to this, I ended up with a bachelor in Scandinavian studies.

But I consciously worked towards my interest in German linguistics at that time. I learned German, and indeed moved to Germany to study linguistics later on. I earned a master's degree in Speech and Language Processing (2019, with a focus on computational linguistics) and a master's degree in General Linguistics (2020, with a focus on formal semantics) from the University of Konstanz. While my huge passion for the German language still lasts until today (and keeps growing stronger every day), it turned out at a later time that NLP / software development fascinate me much more than German linguistics.

Why? I'd say that it comes from my personal preference for hands-on things. In my student life, I gradually found out that I prefer implementing tangible solutions, where I can see an immediate and deterministic outcome on whether it works or not: I see a program or an application growing up from my efforts, and I know that I can apply it to real-world data to solve specific problems - this is what I really enjoy. During my master's study, I worked quite a lot on formal methods for linguistics (formal semantics and formal pragmatics), a very theoretical field that uses mathematical methods (predicate logic, set theory, lambda calculus, etc.) to analyze how meanings are composed from words and phrases. While it was undoubtedly interesting to learn all the methodologies, I struggled quite a few times with the question "ok, now we have a formal model for phenomenon X, but how can this be useful for the real world?" ¹

Then, starting from the later part of my Master's studies, I began working more on NLP. Afterward, I also wrote my dissertation in this field. During this time, I also developed a strong interest in software development: I am not just interested in programming loose scripts of data science solutions, but also implementing them into well-designed applications. Since then, I have felt every day that I am doing the right thing that I truly enjoy.

(¹ A small disclaimer: this is not to be interpreted in the sense that I am anti-intellectual and against fundamental research for which one does not find an immediate application. It's just a personal preference for conducting applied research over fundamental research. As someone who has worked in research for four years, I see the merit of fundamental research, and my forthcoming dissertation will demonstrate how NLP applications can benefit from many well-studied phenomena in theoretical linguistics. That said, I do have a somewhat radical critique of certain work in formal linguistics that use syntax trees with anecdotally defined node labels or overcomplicated lambda calculus formulars: I neither see a possibility of implementing them into computer applications to facilitate artificial intelligence, nor do I believe they are really explaining how humans understand language - I mean, my brain probably does not process a sentence in terms of possible worlds or modal base or any other such machineries that are so prevalent in formal semantics. While mathematization is useful, over-mathematization is nothing but pseudo-science.)

Q: How to pronounce your name (Qi)?
A: Pretty simple! Like "chee" in "cheetah".

Q: Where are you from?
A: Hohhot / hūhéhàotè / Хөх хот - somewhere at the Mongolian steppe. 😉