Implement latest algorithms from research papers for model compression and apply to latest model architectures such as diffusion models, large language models etc.
Set up training jobs, datasets, evaluation, performance benchmarking pipelines.
Applying training time and post training compression techniques
Understanding HW capabilities and incorporating those in optimization algorithm design / enhancement
Build upon the latest research to create new algorithms and invent new ways of applying compression to deep learning models from different domains.
Keep up with the latest AI research and collaborate with diverse teams, both internal and external to Apple, including researchers, hardware architects, and software engineers, to co-develop and implement algorithms customized for Apple hardware.
Run detailed experiments and ablation studies to profile algorithms on various models, tasks, across different model sizes.
Improving model optimization documentation, writing tutorials and guides
