Author

Kang Gyeonghun

“Study hard what interests you the most in the most undisciplined, irreverent and original manner possible.” (Richard Feynmann)

About

mypic

My name is Kang Gyeonghun (Korean: 강경훈, Pronunciation: Khane-gi-yeong-hoon). I am currently an undergraduate student in Yonsei University, Korea. I am majoring in Economics but I was disenchanted with that subject way back, and had dreamed of being an professional investor. I spent most of my eight semesters of undergraduate course studying corporate finance, accounting, portfolio strategy and such, along with an internship in a local private equity.

Gradually I became suspicious of the belief in the industry that “the smart one gets the return”, and were more inclined to believe that “the lucky happens to win”. The book by Nassim N. Taleb, Fooled by Randomness, was the last nail in the coffin to my reservation. All the billionaire star investors I revered looked as if a few out of enumerous gamblers in a coin toss game, who happened to have hundreads of heads in a row. I became more intrigued in the nature of the gamble itself, and that led me into studying Probability and Statistics.

Unfortunately, by the time I found Statistics I had already finished all the regular eight undergraduate semesters, so I had to do extra semesters, and this fall, 2020 semester would be my eleventh. My goal is to continue study Statstics to finish Master’s degree course and apply for Ph.D program, so I am actively preparing and looking for Masters program in Korea or else.

Interest

I am mostly interested in building a probabilistic model that, albeit limited, succintly descirbes generating process of the data. A recent paper I read about the probability model for election forensics, developed by professor Walter R. Mebane sparked my excitement in building and learning a model that enables complex inference. For learning algorithm, I am fairly convinced that Bayesian approach where one gets to inspect the distribution of parameters and models provides a reasonable answers.

Personally, I hold reservation on the widely accepted and used NHST procedure, partly due to my experience of reading many papers about investment strategies where each reached different conclusion, all statistically significant, based on the identical data. I am more convinced that decision making might as well be based on the probability of hypotheses, not by inspecting how rare a sample is under one particular narrow assumption. I believe that the gap between $p(H_0\mid D)$ and $p(D \mid H_0)$ is fairly discussed in the paper by Sellke et al.

Apart from inference where model interpretability matters, I believe that nonparametric methods (e.g. neural networks) are more suitable for prediction when the prediction accuracy is the primal focus.

Experiences

  • Head of Study Prep Team, Expanded Statistics Club (undergrad students' club), Yonsei Univ (Jan 2020 ~ June 2020)

    • Prepared and taught a course on the book [Introduction to Statistical Learning] (Jan 2020) (Youtube (KOR)) (GIthub repo)
    • Prepared and taught a course “Quick Recap of Linear Algebra” (Youtube (KOR))
    • Orchestraed a semester course on the book [Pattern Recognition and Machine Learning] (Mar~june 2020) (Github repo)
    • Now teaching summer course about Bayesian Machine Learning (Aug 2020)

  • Head of Study Prep Team, Head of Investment Team, Yonsei Investment Group (undergrad students' club), Yonsei Univ (Sept 2018 ~ June 2019)

    • Prepared and taught a course “Financial Statetments Analysis for Equity Investment” (Sept 2018 ~ Dec 2018)
    • Prepared and taught a course “Builing Discounted Cashflow Model with Excel” (Jan 2019)
    • Managed 10 mil KRW YIG nvestment fund (Mar 2019 ~ June 2019)

  • Internship at Aarden Partners, Seoul (Apr 2018 ~ July 2018)

    • Local private equity managed by ex investment bankers from Morgan Stanley HK, JP Morgan Seoul with AUM of c. 100 bill KRW at the moment
    • Due to shortage of staff, worked as a defacto junior finanacial analyst; built in-house spreadsheet models, attended and organized meetings with local and foreign clients

  • Air Force Intepreter, intelligence unit, ROKAF Operational Command (June 2014 ~ May 2016)

    • Interpreted for field-level meetings between ROKAF and USAF intel analysts
    • Interpreted for working group conferences on project of ROKAF and DAPA acquiring vital intelligence asset from USAF and foreign private contractor.
    • Traslated educational materials and reports for ROKAF intel analysts