Multimodal AI Engineer — Audio, Video & Speech Pipelines

Meta; Menlo Park (CA); On-site; USD 154,000 - 217,000; Full time

Benefits offered by this job

Equity

Bonus

Benefits

Job summary

Meta’s Applied AI organization seeks a Software Engineer for Multimedia & Multimodal AI to own data production pipelines across image, video, audio, and speech modalities. You will shape what data is produced, how it’s measured, and ensure scalable, high-quality evaluation that informs model development.

You will build end-to-end data workflows, reproduce state-of-the-art research, and mentor engineers while advancing generative and representation models for speech, sound, and music.

Qualifications

• 3+ years building ML systems in production or research.
• Strong Python and PyTorch skills.
• Experience with speech, audio, or music ML (ASR, TTS, codecs, MIR, self-supervised audio, or audio generative modeling).
• Experience with large-scale data pipelines and distributed training.
• Track record of turning research ideas into working, measurable systems.

Responsibilities

• Design and build agentic workflows and pipelines to automate data production and scale output beyond manual authoring.
• Design and own data pipelines: ingestion, filtering, pseudo-labeling and captioning with attribute classifiers, and provenance tracking for audio corpora.
• Build evaluation infrastructure: objective metrics, human listening-test pipelines, and correlation analysis that ties metrics together.
• Improve training efficiency and reliability: distributed training, GPU utilization, codec and tokenizer retraining, experiment management.
• Reproduce and extend state-of-the-art research: implement new methods from papers into our codebases and run rigorous ablations.
• Mentor engineers on the team, contribute to hiring and onboarding, and raise the bar on evaluation and quality practice.
• Build and train generative and representation models for speech, sound, and music — including text-, audio-, and video-conditioned generation, infilling, editing, and style transfer.

Skills

Python

PyTorch

ML systems

Data pipelines

Speech/Audio/Music ML

Research-to-production

Job description

Meta’s Applied AI organization seeks a Software Engineer for Multimedia & Multimodal AI to own data production pipelines across image, video, audio, and speech modalities. You will shape what data is produced, how it’s measured, and ensure scalable, high-quality evaluation that informs model development.

You will build end-to-end data workflows, reproduce state-of-the-art research, and mentor engineers while advancing generative and representation models for speech, sound, and music.