I am leading the NLP team within the applied deep learning research group. My team focuses on building large language models (LLMs) / large Multi-Modal models, application of these models to downstream tasks, and developing new techniques to improve accuracies and performance of generative models. As part of these efforts, we have built Megatron-LM, a framework for efficiently training LLMs at scale and used it to train several large scale models such as Megatron-Turing NLG with 530 billions of parameters.