Adobe Firefly is looking for a senior GenAi applied researcher and solutions architect to architect the training data ecosystem that powers their generative AI models for content synthesis and editing.
Requirements
- Proficient in Python and PyTorch.
- Research or industry experience in training Generative AI models (pre-training and/or post-training) in at least one of the following modalities: image, video, 3D, or audio.
- Expertise in large-scale model training and optimization, including data curation, distributed training, and memory-efficient techniques.
- Experience with post-training techniques such as fine-tuning, alignment or distillation.
- Proven ability in building, deploying, measuring, and maintaining large-scale generative models (e.g., GANs, diffusion models, Transformers).
- Proven ability in building large-scale distributed ML pipelines focusing on Generative Ai.
- In-depth knowledge of machine learning algorithms and their applications to business problems with focus on Generative AI specially image and video.
Responsibilities
- Conduct pioneering research and development in Generative AI, LLMs, LMMs, and reinforcement learning
- Develop and deploy novel generative AI technologies to Adobe Products
- Collaborate with world-class researchers and ML engineers to bring research ideas to production
- Work closely with data scientists, engineers, researchers, and product managers to build AI/ML solutions on the training data ecosystem to deliver a robust, scalable, and efficient data pipelines for the entire training data lifecycle.
- Complete the comprehensive strategy for acquiring, processing, curating, annotating, versioning, and ensuring the quality of large-scale training datasets for Adobe Firefly.
- Work across organizational boundaries to align priorities and drive projects forward.
- Evaluate and integrate new tools, technologies, and methodologies to continuously improve the training data infrastructure, workflows, and team productivity.
Other
- Masters or Ph.D. in Computer Science, Data Science, Engineering, AI/ML, or a related technical field.
- Strong publication record in reinforcement learning, LLMs, and LMMs
- Experience of working with product teams on technology transfers
- Experience with MLOps principles and tools, particularly those focused on data management, data versioning (e.g., DVC), and experiment tracking for ML.
- A track record of contributions to open-source data tools, relevant academic publications, or patents in the data management space.
