Frank Zalkow
Senior Applied Scientist at Microsoft
Greater Nuremberg Metropolitan Area Germany
Summary
👤
Senior🎓
Top SchoolFrank Zalkow is a Senior Applied Scientist with 11 years of experience at the intersection of music information retrieval and speech synthesis, now applying deep learning at Microsoft to bring research-grade audio and TTS technologies into products. He combines a PhD in engineering with a rare blend of music informatics training and hands-on signal-processing expertise, evidenced by contributions to key open-source audio libraries like librosa and music21. His work improves numerical stability and real-world interoperability—e.g., refactoring IIR filters to SOS form and fixing MusicXML clef handling—while also strengthening ML tooling through test automation in scikit-learn. Based in the Nuremberg area, he is fluent in both academic research and production deployment, consistently translating nuanced audio research into robust, test-covered implementations.
11 years of coding experience
4 years of employment as a software developer
Master of Arts - MA Music Informatics, Master of Arts - MA Music Informatics at Hochschule für Musik KarlsruheDoktor (Ph.D.) Engineering, Doktor (Ph.D.) Engineering at FAU Erlangen-Nürnberg
Stackoverflow
3,890reputation
697kreached
72answers
3questions
signal-processing
top-5%
scipy
top-5%
audio
top-1%
python
top-5%
numpy
top-5%
Github Skills (31)
unit-testing10
scipy10
dspic10
python10
testing10
scikit10
audio10
filter10
scikit-learn10
computer-music10
musicxml10
librosa10
dspace10
test-automation10
data-science9
Programming languages (9)
Github contributions (5)
librosa/librosa
Nov 2018 - May 2021
Python library for audio and music analysis
Role in this project:
Back-end Developer
Back-end DeveloperContributions:20 commits, 8 PRs, 38 comments in 2 years 5 months
Contributions summary:Frank primarily contributed to the `librosa` library, focusing on improvements related to IIR filter design and implementation. They refactored filter design using second-order sections (SOS) for increased stability, addressing potential numerical issues. Furthermore, the user implemented a new `flayout` argument to provide flexibility, allowing selection between `ba` and `sos` filter layouts. They also fixed a hopsize issue for IIRT and added unit tests and padding fixes.
python-librarydtwpythonlibrosaaudio
cuthbertLab/music21
May 2015 - Jul 2019
music21 is a Toolkit for Computational Musicology
Role in this project:Back-end Developer & Test Automation Engineer
Back-end Developer & Test Automation EngineerContributions:29 commits, 12 PRs, 17 comments in 4 years 2 months
Contributions summary:Frank primarily focused on enhancing the music21 toolkit, specifically regarding MusicXML import and export capabilities. Their contributions involved implementing mid-measure clef handling and refining the MusicXML conversion process. They also added and extended test files to validate these new features, including checks for clef offsets and multi-stave examples, thus ensuring functionality. Furthermore, they improved the accuracy of enharmonic note naming.
dtwpythonmusicologymusic21music
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.