Taehun Kim

Microsoft Corp.
Audio Communication Group, TU Berlin

- taehun.kim2 AT gmail.com


>professional experiences





>Awards and grants



>research interests<

I am interested in:




Proceedings (peer-reviewed)

Technical reports




Microsoft Corp, Software Development Engineer
Feb. 2016 ~ current

Working for objective-c projects in the Berlin office

MagicInternet GmbH (subsidiary of ProSiebenSat1 Media SE), Senior Mac/iOS Developer
Nov. 2015 ~ Jan. 2016

Developed an App for a new TV experience on iPad from Scratch. Pure Swift project. MVVM + Reactive Programming.

Heap GmbH, Co-Founder (iOS/Data Science)
Oct. 2014 ~ Nov. 2015

Developed highly dynamic mobile media platform for creation and sharing of digital content on mobile devices with intuitive touch interfaces. I am responsible for developing iOS client and its production pipeline including Continuous Delivery and Test Automation. I am also responsible to provide data analysis for understanding our community to help business intelligence.
Heap for iOS

zehndetails GmbH, Head of Frontend R&D
Dec. 2013 ~ Sep. 2014

The company aims at developing an interactive system to discover digital conference posters with intuitive touch interfaces. We explored various touch actions like pinching and dragging to provide better interface to discover large number of scientific conference posters. The system also can present videos as scientific evidences, which was difficult with traditional paper posters. I was responsible for developing interactive iPad app, which also drives external TV as the second screen for the content presentation.
Das Terminal

Smart Mobile Factory GmbH, iOS Developer
Aug. 2011 ~ Nov. 2013

Developing iOS Apps for commercial purposes (Working student).
Selected projects: P.M. Magazin, Wheelmap, Artmapp, WhatWasHere (You can download them from Apple App Store)

Student Project @ the University of Tokyo
Apr. 2009 ~ March 2010

I developed iPhone Apps in the 1-year educational program jointly operated by Japanese Government, the University of Tokyo, National Institute of Informatics Japan and Keio University. Me and 2 other students made a team and we aimed at developing mobile applications for supporting short-distance human communication. In the project we developed Name Card, which supports exchanging contacts information via Bluetooth, and CanvasLive!, which provides a large shared virtual canvas for multiple mobile devices.

Video: NameCard
Video: CanvasLive! #1
Video: CanvasLive! #2


Statistical Synthesis and Analysis of Music Performance
May 2011 ~ current

As a Ph.D. candidate at Audio Communication Group, TU Berlin, I am working on developing statistical models of expressive music performance with machine learning. This research would be useful for developing diverse music intelligent systems, such as content-based music search engine, music navigation and expressive sound synthesis systems. Furthermore, the research will provide a new approach for music interpretation studies with statistical methods.


POLYHYMNIA: a fully automated piano performance system
Apr. 2009 ~ March 2011

This is the main achievement during my Master's course. The system is able to automatically render a given piano piece with a certain performance style. Statistical models of style-aware musical expressions are trained with human expressive piano performances. It is probably the first automatic piano performance system, which can learn polyphonic expression in piano music. Subjective listening tests indicate that the rendered performances sounded like human piano performances, and individual performance styles could be identified.


ORPHEUS: automatic song composition with Japanese lyrics,
Sep. 2008 ~ March 2009

ORPHEUS automatically composes a song with Japanese prosody information, and render it with artificial singing voices. I worked on improving the system by adding the ability of composing songs with musical structures and proposed a method based on combining pre-defined rhythms, chords and accompaniment patterns. The system is currently maintained by SOUND lab. of the University of Tokyo.





Here is private area. Only for authorized persons.

개인 공간입니다. 허락된 사용자만 접근이 허용됩니다.