Applied Scientist - Generative AI Models

About Adobe and the Role

Adobe is dedicated to revolutionizing the digital experience landscape, empowering users ranging from emerging creatives to global enterprises with tools to create compelling images, videos, and applications. We are committed to fostering an inclusive environment where innovation thrives, and everyone has the opportunity to contribute groundbreaking ideas.

The Adobe Firefly ASML team is seeking research scientists and engineers passionate about conditional generation and editing of large-scale generative AI models, focusing especially on images and videos. The goal is to push the boundaries of generative AI while ensuring top quality and fine control in these models.

This role significantly impacts the advancement of Adobe's image and video generation systems, enabling the next wave of creative tools used by millions. You will join a distinguished team of applied researchers and engineers, collaborating extensively to innovate and implement cutting-edge AI technology worldwide.

Key Responsibilities

• Lead the design, development, and assessment of mid-training methodologies that enhance editing capabilities of multi-modal generative models for image, video, and audio.
• Manage specific components within the mid-training framework, such as image-to-image and instruction-based editing, conducting experiments to validate hypotheses and detect quality deficits.
• Create and oversee large-scale captioning systems and support vision-language model fine-tuning efforts to refine multimodal comprehension across visual and auditory data.
• Develop scalable processes for data curation, enhance data quality, and manage distributed training operations by applying latest research in diffusion modeling and large-scale training.
• Work collaboratively with research, data analysis, evaluation, infrastructure, and training teams, contributing insights and detailed documentation on experiments, datasets, and training methodologies.

Qualifications and Skills

• Master’s or Ph.D. in Computer Science, Machine Learning, or a related discipline.
• Deep expertise in contemporary generative AI architectures, particularly diffusion models, along with knowledge of conditional generation and editing techniques for images, videos, and audio content.
• Experience building machine learning models using state-of-the-art deep learning frameworks such as PyTorch, including large-scale and distributed training scenarios.
• Research or practical exposure to vision-language model fine-tuning concentrating on image, video, and audio understanding, adapting pretrained models for downstream applications.
• Familiarity with automated captioning pipelines for handling large datasets involving image, video, and audio modalities.
• Strong programming and prototyping skills in Python and PyTorch.
• Excellent communication abilities and effectiveness in cross-functional team collaboration.

About Adobe

Adobe leads the market by delivering innovative creative and digital experience software including Acrobat Studio, Adobe Express, Firefly, Creative Cloud, and more. Supported by over 30,000 employees worldwide, Adobe fosters a culture that values creativity, equality, and impact.

Commitment to Diversity and Inclusion

Adobe is an Equal Employment Opportunity employer committed to diversity without discrimination on any protected characteristics. We ensure accessibility in recruiting and offer accommodations to candidates with special needs.

Interview and AI Usage Guidelines

Adobe’s interview process aims to reflect individual skillsets fairly. Use of AI or recording software during live interviews is strictly regulated and only permitted with prior approval as a reasonable accommodation.

Compensation

The salary range for this position in the U.S. varies between $120,700 and $238,600 annually, depending on location and candidate experience. Specifically, in California, the range is $164,800 to $238,600. Compensation models vary by role type and may include short-term incentives and long-term equity awards.

Legal Notices

Adobe complies with local regulations such as "fair chance" employment acts and state-specific application policies. Massachusetts law prohibits the use of lie detector tests for employment purposes.