Principal Scientist - Data Pipeline Engineer

Architect and scale multimodal data processing pipelines that power Adobe Firefly’s multimodal foundation models (image, video, and audio). Build distributed, GPU-accelerated systems that transform billions of assets into training-ready data at scale. Optimize ingestion, processing, and delivery by eliminating bottlenecks, improving inference throughput, and designing scalable storage, indexing, and serving infrastructure. Partner with modeling teams to drive data curation and training outcomes.