Research Scientist, Speech/Audio ML — Next-Gen Audio Models

Ifm Us; Paris; Sur place; EUR 90 000 - 130 000; Plein temps

Résumé du poste

Ifm Us is seeking a Research Scientist specialized in speech and audio machine learning to advance end-to-end neural speech models. You will contribute to designing and training state-of-the-art architectures, moving beyond cascaded systems toward native audio-to-audio multimodal solutions.

You will collaborate with researchers, data scientists, and engineers to create core IP, publish guided research, and drive impactful results in AI development.

Qualifications

• PhD or MSc in Computer Science with a focus on Deep Learning, Signal Processing, or Computational Linguistics.
• Record of research with publications in top venues (NeurIPS, ICLR, ICASSP, Interspeech).

Responsabilités

• Architect and train end-to-end neural speech models with SoTA performance.
• Develop core intellectual property for native audio-to-audio multimodal architectures.
• Conduct large-scale experiments and ablations to improve model design and tokenization.
• Evaluate models with objective metrics (WER, MCD) and MOS-based human judgments.

Connaissances

Speech/audio ML

Neural architectures

Model evaluation

Research collaboration

Formation

PhD or MSc in CS/DL/Signal Processing

Description du poste

Ifm Us is seeking a Research Scientist specialized in speech and audio machine learning to advance end-to-end neural speech models. You will contribute to designing and training state-of-the-art architectures, moving beyond cascaded systems toward native audio-to-audio multimodal solutions.

You will collaborate with researchers, data scientists, and engineers to create core IP, publish guided research, and drive impactful results in AI development.