About the job
Exciting opportunity for a Research Engineer to drive innovation in large-scale generative AI model optimization at Adobe Research. Collaborate with top scientists and engineers to develop high-performance ML solutions, impact millions of users, and shape the future of creative tools. Join us to make a real impact in AI and next-generation technology.
Responsibilities
Play a key collaborative role in ambitious research projects; Be valued as a specialist in your domain of expertise; Build innovative tool that enables users to explore their creative potential; Contribute to existing Adobe tools as well as completely new applications; Impact products that are used by tens of millions of people; Learn from your peers and grow into new opportunities
Qualifications
Minimum
Passion for model optimization/compression and high-performance computing; Solid deep learning skills, including practical experience in computer vision/natural language processing; Knowledgeable about the current state of the art in ML efficiency; Experience in improving the efficiency of mid- to large-scale ML models; Software engineering expertise; Proficiency with Python and ML libraries like PyTorch, TensorFlow, JAX, or similar; Strong communication and collaboration skills; Ph.D. /Master's degree in Computer Science or a related field, or 3 years of industry experience
Preferred
Knowledge of state-of-the-art machine learning methods for large scale multimodal models; Experience on diffusion model optimization, neural network pruning/knowledge distillation/quantization/architecture search, sub-quadratic attention optimization, efficient architecture design and on-device ML; Experience with sparse mixture of experts and related techniques; Experience running ML models within a deployment environment (using TensorRT, AITemplate, CoreML, WinML, TensorFlow Lite, ONNXRuntime or similar); Hands-on experience in designing ML models between different platforms (Cloud, mobile, in-browser, etc.); Knowledge of design techniques for mobile-friendly ML models; Proficiency with C++