Conducted performance assessments for diverse accelerators across
four Texas A&M University supercomputers.
Launched job execution on 15 Slurm cluster nodes utilizing Singularity
images to support Large Language Models such as BERT and GPT-2.
Utilized three distinct distributed deep learning strategies: Data
Parallelism, Model Parallelism, and Tensor Parallelism.
Executed comprehensive scaling studies (strong scaling and weak
scaling) on 30 GPUs across single and multi-node configurations.
F
FactSet
Software Engineer
Hyderabad, TG, India
June 2021 - present
Fun Fact
Right from my childhood, I love Math. In my childhood, I stood in 10th place out of 10000 people in a Mathematics Olympiad exam.