Master Thesis – Core Tech Media
Core
Researching machine learning techniques for audio classification, model compression, synthetic data generation, and speaker anonymization for edge devices and intercoms.
Role type
Master's thesis researcher (Computer Science/Applied Mathematics)
Builds
Audio classification models, synthetic audio datasets, energy-efficient inference pipelines, and speaker anonymization systems
Domain
Audio processing, Machine Learning, Embedded Systems, Signal Processing
Deliverable
research
Required skills
C/Python, Machine Learning/AI, Signal Processing
Preferred skills
Knowledge distillation, Model compression (quantization/pruning), Generative models, Synthetic data generation, Speaker anonymization, Embedded hardware optimization
Technologies
pyroomacoustics, AudioLDM, AudioGen, Stable Audio, McAdams coefficient
Responsibilities
Design and implement ML models for audio tasks, evaluate synthetic data strategies, benchmark energy consumption on edge hardware, develop real-time anonymization pipelines
Seniority
Master's student (30hp)
