Matthias Leimeister
Senior Manager, Machine Learning at Sonos, Inc.
Quedlinburg, Saxony-Anhalt, Germany
Summary
🤩
Rockstar🎓
Top SchoolMatthias Leimeister is a Senior Manager in Machine Learning with 11 years of experience building audio and language ML systems, currently leading Sonos’ Audio Machine Learning team responsible for far-field ASR, wakeword detection, audio front-end processing and large-scale domain adaptation. He combines deep signal-processing roots from research and DSP roles (Native Instruments, Fraunhofer) with hands-on ML engineering and productionisation across startups and open-source projects. At Rasa he contributed tests and performance optimizations to a prominent conversational AI framework, showing attention to robustness and reproducibility in ML pipelines. Known for shipping end-to-end solutions—from room simulation and acoustic echo cancellation to scalable model deployment—he balances scientific rigor with product delivery. His background in mathematics and media technology underpins a pragmatic approach to complex audio problems and clever engineering trade-offs.
11 years of coding experience
15 years of employment as a software developer
Master of Science (M.Sc.) Medientechnologie, Master of Science (M.Sc.) Medientechnologie at Technische Universität IlmenauDiplom-Mathematiker Mathematik, Diplom-Mathematiker Mathematik at Heidelberg University
English, German, French
Github Skills (10)
unit-testing10
nlu10
machine-learning10
nlp10
tensorflow10
pytest10
rasa10
python10
bot-framework9
conversational-ai8
Programming languages (2)
Github contributions (5)
RasaHQ/rasa
Nov 2021 - Feb 2022
💬 Open source machine learning framework to automate text- and voice-based conversations: NLU, dialogue management, connect to Slack, Facebook, and more - Create chatbots and voice assistants
Role in this project:
ML Engineer & QA Engineer / Test Automation Engineer
ML Engineer & QA Engineer / Test Automation EngineerContributions:27 reviews, 58 commits, 18 PRs in 2 months
Contributions summary:Matthias contributed to the Rasa project by implementing unit tests for the DIETClassifier and GraphNode, demonstrating expertise in testing machine learning components. They also refactored code related to configuration management and the `override_defaults` function. Furthermore, they addressed a cross-validation issue with lookup tables within the NLU pipeline, implementing a unit test to trigger the problem and providing a solution. The user also optimized the model prediction using `tf.function`.
nlupythonbotspeech-recognitionbotkit
mleimeister/SegmentationCNN
May 2016 - May 2021
Music segmentation using convolutional neural networks.
Contributions:17 commits, 3 PRs, 16 pushes in 5 years
deep-learningconvolutionalneural-networkstensorflowconvolutional-neural-networks
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.