pythonp is a simple utility script that helps you using python on the command line. Basically, it's a python -c command with a handy print function p. You can do something like this with it.

# Remove extensions from files which end with .txt
$ ls | pythonp -e "if l.endswith('.txt'): p('mv', l, l[:-4])" | sh

  By design, it avoids adding much sugar to the good old python -c. It introduces no magic except for a few preprocessings and handy global variables. The goal of this project is to empower users who already know the language and become a part of some major python implementations in the end, without remaining as a standalone package. You can install it via pip install pythonp


  LecRec is a cloud-based automatic sound record and transcription service for iOS/Android platforms created at Open Source Hackathon, KIISE(한국정보과학회), January 2016. I pitched my idea, recruited the team on the spot, successfully led project and made the prototype version of LecRec in three days at this event. This project was awarded the grand prize(최우수상) for its promising usability. The service has not been released in the market and is not maintained any more now that the members started working for other companies. If you still think this project is really interesting and are willing to support it, you may have to contact me to revive it.

official homepage for the event

Instagram Crawler

  When you search for a picture or song you google it with a few words of your choice. How about for thousands, millions or even billions of them? You may not need this much data now but some people do. AI specialists. Because the more data you have, the smater AI you get. But it's a lot of data and you definitely need agents who can do dirty jobs on behalf of you. Web crawlers(programmers) create what are called spiders and make them scrape data on the web. It's a bit like a maze game because everytime a spider runs into a deadend it backtracks

  I developed an Instgram crawler and it could crawl millions of posts in a day. (The illustration is there only to help your understanding).

For more information on the project

Continuous Speech Recognizer

  Implemented Viterbi algorithm for continuous speech recognition as well as constructing hidden Markov model (HMM) for acoustic model probability with bigram for language model probability; an input speech is given as a vector sequence to be parsed.

Machine Learning 2015, Korea University