About

Hi there! I’m Kelly :)

I’m trying to figure out how I can help usher in humanity’s coolest era yet (all the while making sure that we don’t kill ourselves during the process). I’m currently physically based in Oxford 75% of the time and Hong Kong 25% of the time (though I’m always spiritually in some alternate fantasy-cyberpunk version of Edo Japan).

Here are some things that I’m focusing on learning right now:

  • Leveling up in terms of technical skills/knowledge
    • Programming in Python
    • Machine Learning
    • Statistics (I’m being spoon-fed frequentist stats but I’m also being taught on the side by a die-hard bayesian whom I love v v much)
    • Linear Algebra
  • Reading about AI governance, strategy + diplomacy - especially topics concerning China
    • This is a lot less ‘focused’ than it sounds as it involves looking outside of AI (e.g. historical case studies of technological analogues, some meta-science: mega-projects, funding and technological progress…)
  • Picking up Japanese again (something I’ve neglected in the past 3 years)
    • I would LOVE to live in Japan at some point! Sapporo would be amazing, as would Kyoto. Saitama seems to have a lot of capybaras though which makes it v v appealing to me
  • Misc: ‘jamming’ (i.e. improv dancing to trashy music) - mainly during work/study breaks, cooking really good Japanese food (esp rice bowls!)

Here are some things that I’ve gotten up to in the last ~3 years:

  • Read a lot about Linguistics
    • Areas that I found the most interesting: syntax (! !! !!!), semantics + pragmatics, language acquisition, cognitive neuroscience of language, language acquisition (first and second) → I spot a common theme - do you? Phonetics was super duper fun too, but I didn’t find it as gripping as the other sub-disciplines. My only regret was that I didn’t learn more about Linguistic typology.
    • I’ve also spent a sizable amount of time doing Linguistics-related activities that don’t entail actual reading: drawing trees, making weird noises, analysing the speech peculiarities of others… I also gave some Linguistics tutoring lessons to high schoolers hoping to study the subject at uni.
  • Spent a summer working at an early-stage startup
    • My primary tasks included 1) designing + conducting a user study → analysing and presenting my findings → translating these into concrete business decisions 2) designing + implementing the marketing strategy of an early-stage startup
  • Constructed some NLP datasets
    • Multi-SimLex - a “large-scale multilingual resource for lexical semantics”
    • PoliteInContext: a contextualised politeness dataset inspired by Terkourafi’s frame-based approach to politeness → this was my undergrad thesis so you won’t be able to find this online :)
  • Helped organise two Hackathons - always had so much fun during the actual event!
  • More recently: did a research project at the intersection of China and AI strategy for a longtermist think-tank in China
    • Actually (sort of) picked up the basics on navigating Chinese cyberspace. Also managed to put my very rusty Chinese reading skills to use. The auto-Google-Translate tool was useful for the first time - it saved me lots of time by translating Chinese texts from simplified to traditional (though I should really learn how to read simplified)
  • Danced danced danced (!!!). I experimented with various different styles:
    • Latin dancing
      • chachacha, jive mainly - but have also done some samba, rumba and paso doble
    • Ballroom dancing (waltz and quickstep)
      • I really wanted to learn Tango but it wasn’t in the beginner’s training syllabus. I was hoping of trying it this year but COVID-19 happened :(
    • Commercial +/ hip hop
      • Since COVID, my main source of dance choreographies has been 1 Million Dance Studio videos. I’ve v sporadically learnt random 1 Million choreos since high school, but have amped up the frequency of this recently. The studio has started releasing a lot more tutorials than they’ve used to but most of them aren’t free — my typical learning method is finding a mirrored version of the dance on Youtube and slowing it down to 0.5.

In a previous life, I wrote a lot of fiction, tried learning a lot of languages and regularly jumped around on ice with very thin blades attached to my feet (with all due seriousness: blades on figure skates are only ~4mm thick which is actually insane).

Feel free to reach out to me at lkw33cam [at] gmail [dot] com - I love meeting new people! I’m absolutely abysmal at replying messages/emails so please expect a delay in my response (it’s nothing personal!).

Side note: if you’re looking for my institutional affiliations, you’ll find them on my LinkedIn Profile (I try to leave conventional signalling to LinkedIn)