Research Scientist - Speech/Audio Machine Learning

On-siteSeniorResearch Scientistposted 10mo ago

✦Role summaryAI-generated

This position focuses on developing and evaluating speech and audio models within a foundation model research lab in Paris. Candidates will collaborate with leading researchers to push the boundaries of audio AI and contribute to transformative industry applications.

Skills required

Speech recognition Audio signal processing Deep learning Neural networks Foundation models PyTorch TensorFlow Research methodology Model evaluation AI Machine learning LLMs

About this role

About the Institute of Foundation Models

We are a dedicated research lab for building, understanding, using, and risk-managing foundation models. Our mandate is to advance research, nurture the next generation of AI builders, and drive transformative contributions to a knowledge-driven economy.

As part of our team, you’ll have the opportunity to work on the core of cutting-edge foundation model training, alongside world-class researchers, data scientists, and engineers, tackling the most fundamental and impactful challenges in AI development. You will participate in the development of groundbreaking AI solutions that have the potential to reshape entire industries. Strategic and innovative problem-solving skills will be instrumental in establishing MBZUAI as a global hub for high-performance computing in deep learning, driving impactful discoveries that inspire the next generation of AI pioneers.

The Role

As a Research Scientist specializing in speech/audio machine learning, you will contribute to the design and training of SoTA end-to-end neural speech models. You will be responsible for developing the core intellectual property, moving beyond cascaded ASR → TTS systems toward native audio-to-audio multimodal architectures.

Key responsibilities

Architectural Design: Develop novel neural architectures for low-latency speech-to-speech translation and generation (e.g., Diffusion, Flow-matching, Transformer-based audio LLMs).

Loss Function Engineering: Design and implement custom objective functions to optimize prosody (emotions, intelligibility, naturalness).

Experimental Iteration: Conduct large-scale training runs, performing ablation studies on model architecture and tokenization strategies.

Evaluation Frameworks: Establish rigorous internal benchmarks using both objective metrics (WER, MCD) and subjective human-in-the-loop (MOS) testing.