Principal Scientist, Foundation Model Understanding Data Architect

Principal Scientist (P60) – Data and Semantic Architecture

We are seeking a Principal Scientist (P60) to define and lead the data and semantic architecture that enables Adobe Firefly’s next‑generation unified foundation models—models that natively combine understanding, reasoning, and generation within a single system. This role sits within the Applied Science and Machine Learning (ASML) organization and operates at the highest level of technical leadership. You will shape how meaning, intent, and abstraction are represented and learned at foundation‑model scale, influencing multiple generations of models and products.

This is a highly strategic individual contributor role with broad organizational impact.

• Define the multi‑year vision for understanding‑centric data that enables unified comprehension and generation.
• Establish core principles for semantic representation, supervision, and curriculum at foundation‑model scale.
• Identify limitations in existing data paradigms and drive the invention of new approaches.

• Design canonical semantic abstractions that bridge perception‑level signals with high‑level concepts such as intent, reasoning, and compositional structure.
• Lead the evolution of supervision schemas toward rich, model‑interpretable semantic signals.
• Guide the use of multimodal and vision‑language models to bootstrap scalable understanding data.

• Define understanding‑focused data curricula across model scale and training phases.
• Lead ablations to quantify how semantic data choices affect generalization, cont rollability, and robustness.
• Balance organic, synthetic, and model‑generated data to optimize learning efficiency.

• Serve as the technical authority for foundation‑model understanding data across Adobe.
• Influence research, infrastructure, and product roadmaps through principled, data‑driven guidance.
• Mentor senior scientists and architects; set standards for rigor and reproducibility.

• 10+ years of experience in ML, data systems, or AI research, including leadership on large‑scale or foundation models.
• Demonstrated impact shaping data, learning paradigms, or representations for LLMs or multimodal models.
• Deep expertise in semantic representation, reasoning, and abstraction.
• Strong intuition for interactions between data, model architecture, and training dynamics.
• Ability to influence senior technical and business leaders through clear technical judgment.

• In Computer Science, Machine Learning, AI, or a related field strongly preferred.