Satish Pasumarthi
Senior Software Engineer at NVIDIA
Hayward, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSatish Pasumarthi is a senior software engineer with 8 years of experience and a strong focus on machine and deep learning infrastructure, currently optimizing DL workflows at NVIDIA after driving model enablement and distributed training features at AWS. He combines hands-on expertise in TensorFlow and PyTorch with DevOps know-how—building and tuning Docker-based deep learning containers, CI/CD for TF, and enabling heterogeneous clusters on SageMaker, including Trainium support. His background in CAD and chip-design tooling lends him a practical systems mindset that shows up in performance-driven model optimization and distributed training resilience. A versatile contributor and mentor, he has open-source commits to prominent AWS ML repos and a track record of reducing operational friction for large-model training on cloud GPUs/accelerators.
8 years of coding experience
13 years of employment as a software developer
B.Tech Electronics and Communication Engineering, B.Tech Electronics and Communication Engineering at National Institute of Technology WarangalMaster's degree (Thesis) Computer Engineering, Master's degree (Thesis) Computer Engineering at Texas A&M University
English, Telugu, Hindi
Github Skills (25)
docker10
distributed-training10
python10
mpi10
dockers10
sagemaker10
deep-learning10
tensorflow10
aws10
cuda10
async9
asynchronous9
cicd9
asyncio9
machine-learning9
Programming languages (6)
Github contributions (5)
aws/sagemaker-training-toolkit
Oct 2021 - Oct 2022
Train machine learning models within a 🐳 Docker container using 🧠 Amazon SageMaker.
Role in this project:
MLOps Engineer
MLOps EngineerContributions:69 reviews, 10 commits, 35 PRs in 1 year
Contributions summary:Satish primarily focused on improving the reliability and functionality of the SageMaker training toolkit. Key contributions include enhancing logging capabilities for better debugging, enabling custom failure logging, and addressing issues related to shell script execution. The user also made significant changes to support heterogeneous cluster configurations and Trainium instances, reflecting a strong understanding of distributed training environments and infrastructure. This involved updates to the environment setup and torch_distributed support.
sagemakerpythonamazon-sagemakerdeep-learningamazon
aws/deep-learning-containers
Jun 2020 - Aug 2022
AWS Deep Learning Containers are pre-built Docker images that make it easier to run popular deep learning frameworks and tools on AWS.
Role in this project:DevOps Engineer & ML Engineer
DevOps Engineer & ML EngineerContributions:62 reviews, 19 commits, 18 PRs in 2 years 2 months
Contributions summary:Satish's commits primarily involve building and configuring Docker images for TensorFlow serving and training within the AWS Deep Learning Containers. They focused on adapting Dockerfiles for different TensorFlow versions, CUDA versions, and training/inference configurations. Their contributions included installing dependencies, configuring environment variables, and optimizing the images for performance. The user also collaborated with other contributors and contributed to the build process and updated the docker files to include cuda11 and cudnn8.
pytorchsagemakercontainersmxnetserving
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.