Miguel Fernández
MIR & AI Researcher at Klangio – AI for Musicians
Munich, Bavaria, Germany
Summary
👤
Senior🎓
Top SchoolMiguel Fernández is an AI and MIR researcher based in Munich with nine years of experience applying signal processing and deep learning to audio and spatial personalization. He spent five years at Huawei developing deep-learning approaches for singing voice transcription and other MIR tasks, and now builds AI tools for musicians at Klangio. His academic background includes a Master's in Sound and Music Computing and ongoing PhD work in information and telecommunication technologies, grounding his work in both theory and practical systems. Miguel contributes to open-source audio tooling—improving CQT implementations and chroma features in a popular PyTorch-based audio library—highlighting his focus on robust feature extraction. Colleagues describe him as curious and versatile, comfortable moving between embedded firmware testing, QA, and research-grade model development. He combines hands-on engineering with research rigor to deliver audio ML solutions that matter to creators and researchers alike.
9 years of coding experience
1 year of employment as a software developer
University of SevillePh.D. Student Information and telecommunication technologies, Ph.D. Student Information and telecommunication technologies at Universitat Pompeu Fabra
Github Skills (7)
pytorch10
audio-processing10
spectrogram10
python10
scipy9
machine-learning9
neural-network8
Programming languages (4)
Github contributions (5)
KinWaiCheuk/nnAudio
May 2021 - Nov 2021
Audio processing by using pytorch 1D convolution network
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 11 commits, 6 PRs in 6 months
Contributions summary:Miguel primarily contributed to the audio processing library by modifying and improving CQT (Constant-Q Transform) implementations. Their work involved fixing bugs, updating import statements, and adding support for non-fp32 data types. The user also added chroma calculations to the library, demonstrating a focus on feature extraction techniques relevant to audio analysis. Furthermore, the user made changes related to file organization and setup configuration by updating the `setup.py` file with scipy requirements and added tests for the functionality.
pytorchaudio-processingspectrogrammelspectrogramconvolution
migperfer/nnAudio
May 2021 - Mar 2023
Audio processing by using pytorch 1D convolution network
Contributions:2 PRs, 24 pushes, 3 branches in 1 year 11 months
pytorchaudio-processingspectrogramconvolutionaudio
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.