Ingénieur·e R&D IA et traitement du signal audio – transcription de musiques à tempéraments inégaux

About the role

The Institut Mines‑Télécom (IMT‑Atlantique) seeks a research and development engineer specialized in artificial intelligence and audio signal processing to advance automatic transcription of music with unequal temperaments. The 24‑month fixed‑term position is based in Brest and will contribute to the ANATIEBA project.

Key responsibilities

• Adapt multi‑fundamental‑frequency (multi‑f0) estimation methods, especially deep‑learning approaches, for transcribing historic recordings of biniou/bombarde in micro‑tonal scales.
• Design, record, annotate and publish a multi‑track reference dataset of the biniou/bombarde couple in ancient tuning systems.
• Promote research outcomes and maintain scientific links with other consortium tasks (T1, T3, T4) and the traditional music community.

Required profile

• Strong proficiency in scientific programming with Python.
• In‑depth knowledge of music theory and its relationship to musical acoustics.
• Expert instrumental practice of the bombarde, biniou or bagpipe, demonstrated by a degree, professional teaching, or equivalent experience.
• Intimate familiarity with early 20th‑century biniou/bombarde repertoire and historic recordings, as well as published research on ancient scales.
• Fluent French (spoken and written).

Required skills

• Python programming.
• Deep learning applied to audio.
• Audio source separation.
• Non‑negative matrix factorization.
• Audio recording and studio work.

What we offer

• 24‑month fixed‑term contract.
• Collaboration with leading research labs (Lab‑STICC/UBO, LAUM/Le Mans University) and a network of musicians, luthiers, and musicologists.
• Access to unique historical audio archives and the chance to contribute to cultural heritage preservation.
• Opportunities for scientific publications and open‑science data/software dissemination.