Doctoral Researcher in Speech and Language Technology

Aalto University is seeking a motivated Doctoral Researcher to join the COWAMA project, which focuses on content-based watermarking and detection for AI-generated audio. The successful candidate will work under the supervision of Assistant Professor Lauri Juvela at the Department of Information and Communications Engineering. The research involves developing methods to identify the origin and authenticity of generated audio, evaluating robustness against adversarial attacks, and contributing to open-source software. This position offers access to high-performance computing infrastructure and an international collaborative network. Applicants should be passionate about speech AI, deepfake detection, and trustworthy generative technologies.

• Develop and implement methods for detecting the authenticity and origin of AI-generated speech and audio.
• Curate diverse datasets of real and synthetic audio content for model training and evaluation.
• Improve detector baselines for multi-objective tasks including deepfake detection and watermark identification.
• Create explainable and localized detection methods for attribution in mixed audio sources.
• Evaluate model robustness against signal-processing, adversarial, and resynthesis attacks.
• Publish research findings in top-tier journals and conferences.
• Contribute to open-source releases of software, trained models, and reproducible research outputs.