Principal Engineer, Efficient GenAI

Advanced Micro Devices, Inc•San Jose, CA

10d•Hybrid

About The Position

The AI Models and Applications team at AMD is looking for a specialized Principal Engineer who is passionate about enabling innovative and efficient Generative AI training and inference at scale. You will be part of a core team of incredibly talented specialists and work on scaling training and inference for the latest Generative AI models.

Requirements

Deep technical understanding and hands-on experience with the latest Generative AI applications in at least one of the following areas: large language models (LLMs), 3D World and Action Models, or image/video generation models.
Experience training models at scale.
Passion for developing efficient approaches to enable distributed training and inference on AMD devices.
Strong technical expertise in Generative AI model training and inference, with familiarity working with deep learning frameworks such as PyTorch, JAX, vLLM, SGLang, and MuJoCo.
Strong technical expertise in algorithmic innovation for efficient Generative AI applications across both training and inference.
Excellent written, verbal, and presentation skills, with the ability to coordinate effectively both internally and externally.
Several years of experience in AI, deep learning, and related software development.
PhD or master’s degree in computer science, Electrical Engineering, Mathematics, or a related field.

Nice To Haves

Expertise and publications in one or more of the following preferred areas: efficient model architectures, optimized training, innovative parallelism strategies, or low-precision training.
Additional plus if publications have been presented at conferences such as NeurIPS, CVPR, ECCV, ICCV, ICML, or ICLR.
Experience productizing Generative AI models and training foundation models at scale.

Responsibilities

Propose and apply innovative techniques to support both training and inference, including innovative transformer architectures, parallelism strategies for training on large clusters, inference optimization techniques such as speculative decoding, and optimal KV-caching strategies.
Implement novel, efficient architectures for Generative AI models for training and inference and showcase the benefits on AMD platforms.
Work with open-source frameworks and communities (e.g., PyTorch, JAX, vLLM, SGLang) to integrate AMD-optimized models and libraries and publish training recipes.
Collaborate with software and hardware teams to co-optimize end-to-end performance on current and future AMD solutions.
Increase adoption of agentic workflows for optimizing and deploying Generative AI applications at scale on AMD platforms.
Publish and promote your work at external venues, including major conferences.
Collaborate with researchers within AMD and across industry and academia to promote innovation on AMD platforms.