AI Lyrics-to-Melody Alignment Engineer

Reltrix Studio

Remote · $160k – $200k / year · Full-time · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Text-to-song generation can produce melodies. It can produce lyrics. What it consistently fails to do is produce melodies that feel natural to sing with specific lyrics — where the natural stress patterns of the words align with musical accents, where phrase lengths match lyric line lengths, and where the melodic contour feels expressive rather than arbitrary relative to the emotional content of the words. This alignment problem sits at the intersection of music cognition, linguistics, and generative modeling. Natural language has prosodic structure: syllables have stress, phrases have rhythm, and emotional content correlates with prosodic contour. Music has an analogous structure. Lyrics-to-melody alignment is the problem of generating music that resonates with the prosodic structure of specific text. Our music generation platform serves 4,000 independent artists and music supervision clients. You'll lead the lyrics-to-melody alignment research: designing models that extract prosodic structure from lyrics, training alignment models that generate melodically appropriate settings, and building the evaluation framework that measures alignment quality perceptually, not just statistically. Prior publications in music information retrieval or generative audio ML are a strong plus.

Fully remote.

Responsibilities

Design lyrics-to-melody alignment models incorporating syllable stress and prosodic contour
Build lyric prosody extraction pipeline identifying stress patterns, phrase boundaries, and emotional valence
Train alignment models conditioned on lyric prosodic features to generate appropriate melodic contours
Develop perceptual evaluation framework using listening tests and expert musician feedback
Build alignment quality metrics correlating with human naturalness judgments
Publish methodology and contribute to music information retrieval research community

Requirements

6+ years in ML research with audio, music information retrieval, or generative modeling experience
Understanding of music theory, prosody, and text-to-speech prosodic modeling
Strong Python and PyTorch skills with audio signal processing experience
Familiarity with music generation models and evaluation approaches
Research publication record in music AI, audio ML, or adjacent NLP fields preferred