Applied Scientist 3
Core
Refining and compressing large-scale generative AI models for images and videos to improve quality, efficiency, and deployability for Adobe Firefly.
Role type
Senior Applied Scientist / Machine Learning Engineer (Post-training & Distillation)
Builds
Production-ready generative AI models and training pipelines for Adobe Firefly
Domain
Generative AI, Computer Vision, Large Language Models
Deliverable
production ML models
Required skills
Machine learning algorithms, Model distillation techniques, Python programming, AI model training and optimization, Supervised fine-tuning (SFT), Preference optimization (DPO/GRPO), Reward-based learning, Model compression, Inference performance optimization
Preferred skills
Large-scale data pipelines, Adobe products knowledge
Technologies
Python, SFT, DPO, GRPO
Responsibilities
Develop and run distillation pipelines to transfer capabilities from large teacher models into smaller student models; Carry out and refine post-training methods including SFT, DPO, and reward-based learning; Build infrastructure and tools for teacher rollout creation, distillation data pipelines, and training workflows; Carry out experiments aimed at improving model quality, efficiency, and instruction alignment; Collaborate with research scientists to convert research ideas into scalable training pipelines; Evaluate models using automated metrics and human preference signals; Optimize models for deployment efficiency including distillation and inference performance
Seniority
Senior, hands-on IC
