Multimodal AI Engineer: Multimedia Pipelines & Evaluation

Meta; Boise (ID); On-site; USD 154,000 - 217,000; Full time

Job summary

Applied AI (AAI) at Meta focuses on making our AI models best-in-class across multimodal modalities such as image, video, audio and speech. This role centers on designing data pipelines, measurement strategies, and evaluation frameworks that scale beyond text, shaping standards for the team.

You will own end-to-end data production, from data selection to provenance, and drive quality improvements while mentoring engineers and collaborating with research, model-training and engineering partners.

Qualifications

• Bachelor's degree in CS/engineering or equivalent practical experience.
• 6+ years programming in a relevant language or 3+ years with PhD.
• 3+ years building ML systems in production or research.
• Strong Python and PyTorch skills.
• Experience with speech, audio, or music ML (ASR/TTS/genres).
• Experience with large-scale data pipelines and distributed training.
• Proven ability to translate research ideas into working systems.

Responsibilities

• Design and build agentic workflows and pipelines with human-in-the-loop designs.
• Own data pipelines at scale: ingestion, filtering, labeling, provenance tracking.
• Build evaluation infrastructure: objective metrics, human tests, correlation analysis.
• Improve training efficiency: distributed training and resource utilization.
• Reproduce and extend state-of-the-art research in codebases.
• Mentor engineers, contribute to hiring and onboarding, raise quality standards.
• Develop generative and representation models for speech, audio, and music.

Skills

Bachelor's degree

Programming experience

ML systems experience

Python

PyTorch

Speech/Audio/Music ML

Large-scale pipelines

Research-to-production

Education

Bachelor's degree in CS/Engineering

Job description

Applied AI (AAI) at Meta focuses on making our AI models best-in-class across multimodal modalities such as image, video, audio and speech. This role centers on designing data pipelines, measurement strategies, and evaluation frameworks that scale beyond text, shaping standards for the team.

You will own end-to-end data production, from data selection to provenance, and drive quality improvements while mentoring engineers and collaborating with research, model-training and engineering partners.