Karan Jariwala
Software Engineering Manager at Meta
Sunnyvale, California, United States
Summary
🤩
Rockstar🎓
Top SchoolKaran Jariwala is a software engineering manager with 9 years of experience building large-scale deep learning infrastructure and distributed training systems, now leading PyTorch-related platform work at Meta from Sunnyvale. He previously managed GenAI training engine efforts at Databricks after helping design core training libraries at MosaicML and scaling distributed training at AWS SageMaker, where his optimizations delivered substantial throughput and scaling efficiency gains. Karan is a hands-on MLOps/ML engineer and active open-source contributor — notable contributions include improving SageMaker's training toolkit for SMDataParallel and enhancing MosaicML’s Composer and Streaming libraries for robust, efficient model training. He blends production-grade systems engineering with ML research sensibilities, shipping end-to-end solutions from dataset conversion scripts to MPI/EFA performance tuning. Comfortable mentoring cross-functional teams, he pairs a background in embedded systems and electrical engineering with advanced CS research from RIT, enabling him to translate silicon- and cluster-level constraints into pragmatic software designs.
9 years of coding experience
11 years of employment as a software developer
Master’s Degree Computer Science, Master’s Degree Computer Science at Rochester Institute of TechnologyPost-Graduate Diploma in Embedded Systems & Design Computer Software Engineering, Post-Graduate Diploma in Embedded Systems & Design Computer Software Engineering at Centre for Development of Advanced Computing (C-DAC), PuneBachelor of Engineering (B.E.) Electrical Electronics and Communications Engineering, Bachelor of Engineering (B.E.) Electrical Electronics and Communications Engineering at Sarvajanik College of Engineering & TechnologyScience Mathematics chemistry Physics, Science Mathematics chemistry Physics at H.M.B Sardar high school
English, Hindi, Gujarati
Github Skills (25)
pytorch10
docker10
distributed-training10
artificial-neural-networks10
python10
testing10
mpi10
machine-learning10
dockers10
datasets10
sagemaker10
amazon-sagemaker10
deeplearning-ai10
mlops10
deep-learning10
Programming languages (4)
Github contributions (5)
mosaicml/streaming
Sep 2022 - Jan 2023
A Data Streaming Library for Efficient Neural Network Training
Role in this project:
ML Engineer
ML EngineerContributions:9 releases, 883 reviews, 47 commits in 3 months
Contributions summary:Karan contributed to data conversion scripts for the COCO2017 and ADE20K datasets, which are common for object detection and image segmentation tasks. The commits involved adding new conversion scripts, fixing code issues, and restructuring existing code. The user also added pre-commit configurations.
pytorchstreaming-datadeep-learningdatasetstreaming
aws/sagemaker-training-toolkit
Feb 2021 - Apr 2021
Train machine learning models within a 🐳 Docker container using 🧠 Amazon SageMaker.
Role in this project:MLOps Engineer
MLOps EngineerContributions:6 reviews, 5 commits, 5 PRs in 2 months
Contributions summary:Karan focused on enhancing the `sagemaker-training-toolkit` repository, specifically for SMDataParallel training. They implemented features such as custom MPI options and enabled EFA RDMA flags for improved performance. Furthermore, the user addressed connection issues between workers by improving error messages and added configuration options. These changes improve the usability and performance of the training toolkit within the SageMaker environment.
sagemakerpythonamazon-sagemakerdeep-learningamazon
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.