Applied Scientist Multimodal

Generative AI is redefining creativity. Ensuring that these systems are safe controllable and respectful of intellectual property is one of the most important open research challenges in the field.

The Adobe Firefly Applied Science & Machine Learning team is building next-generation multimodal guardrail systems forbuildingsafe and compliantimage video and audio generative models powering . Our goal extends beyond reactiveblocking;we are developing model-level guidance mechanisms that proactively steer generation away from IP-violating concepts while preserving creative intent and usability.

We areseekingaApplied Scientistwith strong multimodal depth and research instincts to help define anddevelop the frontierof IP-aware generative modeling. This role sits at the intersection of generative model alignment multimodal reasoning and large-scale inference systems.

Research Areas You Will Drive

Inference-Time Alignment & Optimization

• Researchand implementinference-time control techniques (guided decoding constrained sampling classifier guidance reward-based steering).

• Optimizelarge multimodal systems for low-latency production-scale deployment without sacrificing alignment quality.

• Identifyand mitigate failure modes in generative pipelines at scale.

Rapid Scientific Experimentation

• Develop and complete rigorous experiments to evaluate trade-offs between creativity fidelity and IP safety.

• Develop new evaluation methodologies and benchmarks for multimodal IP compliance.

• Contribute novel technical insights that may lead to publications or internal intellectual property.

Vision-Language & Multimodal Reasoning

• Advance the use of Vision-Language Models (VLMs) and multimodal foundation models for semantic IP understanding.

• Explore joint reasoning betweenperceptionand generation systems to enable real-time steering.

• Experiment how various techniquese.g.multimodal embeddingscross-attentionmechanismsetc. canbeusedfor safety-aware inference.

Multimodal IP-Aware Generative Modeling

• Develop novel approaches for integrating IP constraints directly into generative model behavior.

• Investigate controllable generation techniques that shift models from post-hoc blocking toward guided alignment-aware synthesis.

• Develop training and fine-tuning strategies that embed guardrail signals into model representations.

WhatYoullDo

• Architect and evolve the Firefly IP Guard pipeline across first-party and third-party generative models.

• Collaborate with research scientists ML engineersandotherkey partnerslike applied ethicslegaletc.to translate scientific advances into deployed systems.

• Communicate with other non-technical collaborators to help drive awarenessabout the implemented technologies.

• Drive end-to-end experimentationfrom hypothesis formulation through model implementation and large-scale evaluation.

• Contribute to the broader research strategy around generative AI alignment and safety within Adobe.

WhatYoullNeed to Succeed

Research & Technical Depth

• PhD or MS in Computer Science Machine Learning AI or related field.

• 5 years of experience in applied ML or generative AI research (industry or academia).

• Strong background in large-scale generative models (diffusion models multimodal transformers autoregressive systems).

• Deep experience with model fine-tuning alignment strategies and representation learning.

• Expertisein Vision-Language Models or multimodal foundation models.

• Proficiencyin Python and modern ML frameworks (e.g.PyTorch) withexperience oftraining and deploying large models.

Scientific & Systems Rigor

• Strong experimental development and statistical evaluation skills.

• Experience analyzing complex failure modes in multimodal systems.

• Understanding oflarge-scale inference systems andproductionML constraints.

• Ability to navigateopen-ended research spaces andidentifyhigh-leverage problems.

AI-Accelerated Development

• Demonstrated ability touseAI coding tools and AI-assisted development workflows to rapidly prototype experiment and scale research impact.

• Comfortoperatingin an AI-augmented development environment using generative tools to increase iteration speed code quality and research throughput.

• Ability to combine scientific rigor with high-velocity execution.

Collaborative Approach

• Experience working in cross-functional research-to-product environments.

• Ability to clearly communicate complex scientific ideas to diverse collaborators.

Preferred Qualifications

• Research contributions in controllable generation alignment AI safety or multimodal learning.

• Publications in leading conferences (CVPR ICCV NeurIPS ICML ICLR SIGGRAPH) or equivalent industry impact.

• Experience deploying generative models to large user bases.

• Background in safety evaluation frameworks explainabilityor adversarial robustness.