Postdoctoral researcher · HfM Weimar

I build tools to explore how music works

I am a computational musicologist, producer, and research software developer. My work grows out of a lifelong fascination with music, computers, and the unexpected things they can do together.

Portrait of Egor Polyakov

It started with a computer—and with having fun

As a child, I was given a ZX Spectrum clone. I was amazed that one small machine could draw, calculate, run games, and even make music, however limited the sound was. Around the same time, I began learning piano and cello and discovered how much I enjoyed improvising and composing. For years, I followed that musical path meticulously and studied classical music almost exclusively.

The connection between those two worlds became clear only much later. What I have always enjoyed most is building a setup—musical, analytical, or technical—then trying it out and discovering an interaction I could not quite see before. Today that might mean turning a score into explorable data, translating a listening question into an experiment, or asking what a machine-learning model can—and cannot—understand about musical features. At heart, I do this work because I have fun doing it. Building accessible tools lets me share that experience with others.

Research and development at HfM Weimar

At the University of Music FRANZ LISZT Weimar, I work on research and development for CAMAT, the Computer-Assisted Music Analysis Toolbox. The DFG-funded project Development of a Comprehensive Cloud-Based Toolbox for Music Score Analysis runs from September 2025 to August 2028, with Prof. Dr. Martin Pfleiderer as principal investigator. We bring MEI, Python, and Jupyter notebooks into musicological research and teaching: editing and checking encodings, connecting notation to facsimiles, extracting musical events, and exploring patterns in scores. Jupyter4NFDI makes these tools accessible through a browser.

The editorial work is part of this, too. In the CAMAT editions, we correct and review optical music recognition from the Bavarian State Library and musiconn.scoresearch, turning it into usable MEI editions and research corpora. Current work draws on Denkmäler deutscher Tonkunst (DdT): volume I/11, Dietrich Buxtehudes Instrumentalwerke, and volumes I/29–30, Instrumentalkonzerte deutscher Meister.

For me, building analytical software and working carefully with the sources belong together. An encoding, explicitly linked back to the facsimile, should offer another way into the source, not replace it or hide the editorial choices made along the way. Whatever we choose to encode, it should remain possible to return to the source and question that reading.

Production, mastering, and live electronics

Alongside research, I have spent much of my working life in the studio. As one half of Two Unknown Guys, I produced minimal and techdub tracks released on Ornithopter Records and Sinergy Networks. As part of Ornithopter’s founding team, I mastered the label’s catalogue, and later releases by bands and underground artists from Leipzig and Berlin.

For concerts and installations, I realised live electronics and stem mastering, most closely with Fabian Russ: premieres at Futurium Berlin, Montforthaus Feldkirch, and Scala Esslingen, several of them interdisciplinary productions with dance, theatre, and video.

Works, releases, and mastering credits

A path through different ideas of music

This was not a neat progression in which one discipline simply led to the next. Encounters with different kinds of music, years of working between them, and a period of doubt all changed the direction of my work.

  1. 01 The classical path

    Piano, cello, improvisation, and composition led me from childhood curiosity to formal studies. For years, I treated the classical tradition as the centre of my musical world.

  2. 02 Beyond the academy

    Love Parade in 2002 and several open-air festivals in 2003–04 profoundly changed that view. Techno, drum and bass, dub, and other forms of popular electronic music made me question the idea of one “true” music.

  3. 03 Programming music

    From around 2008–09, Max/MSP, OpenMusic, Pure Data, and related environments became central to how I worked. I built real-time audio systems, Max for Live devices, and analysis–synthesis setups for composition, performance, and teaching.

  4. 04 Between worlds

    For years I composed contemporary concert music and minimal/techdub in parallel. Alongside my own music, I mastered the Ornithopter Records catalogue as part of its founding team and worked with Fabian Russ on concert and recording projects. Over time I came to see composition as one part of my work rather than the whole of it: producing, realising other people’s music, and building tools became more central.

  5. 05 From real time to research

    My doctorate led me from accessible software and hybrid analysis systems towards MIR and computational musicology. As my focus moved away from real-time audio around 2020, Python gradually replaced patching environments; CAMAT in 2021–22 brought these threads together in open analytical tools and, later, machine learning.

Open tools, reproducible music

I do not think access to music or musical analysis should depend on proprietary software or expensive hardware. Accessibility is a design requirement in my work: I release my tools as open source and keep their methods, assumptions, and workflows inspectable, reproducible, and adaptable.

Reproducibility also concerns the medium itself. I want people to be able to experience or study music with whatever is available: a sophisticated studio system, inexpensive headphones, a small loudspeaker, or even a visual or analytical representation. Music organises time; sound is one way of encountering that organisation, but not the only one.

Selected projects

All projects

CAMAT

MEI editing, facsimile-linked editions, and computational music analysis with Python and Jupyter notebooks.

Visit project

AudioSpylt

Python audio analysis, signal processing, and interactive visualization, inspired by AudioSculpt and OpenMusic.

Visit project
All projects

Recent highlights

All publications
  • 2026 — Presented poster “CAMAT_v2: Cloud-Based Tools for Collaborative Music Score Analysis”, ICCCM26, Würzburg
  • 2026 — Presented “Beyond the Black Box…”, InMusic26, Aalborg
  • 2026 — Published chapter on FFTimbre in Routledge Innovation in Music: Current Research Perspectives
  • In press — Chapter on the beat_it toolbox in Routledge Innovation in Music: Sound and Place
All publications

Research, teaching, and collaboration

For professional enquiries, the most direct route is email.