Israt Nisa
Research Scientist at Meta
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolIsrat Nisa is a research scientist with 10 years of experience building high-performance GenAI inference systems and scalable graph neural network models across GPUs and custom AI silicon. Currently at Meta after a multi-year applied science role at AWS, she blends systems co-design with hands-on CUDA engineering to accelerate sparse and irregular algorithms. Her PhD and research background at The Ohio State University and Berkeley Lab underpin expertise in GPU-optimized sparse kernels, task parallelism, and exascale microbiome analysis pipelines. A notable open-source contributor to DGL, she implemented and optimized CUDA kernels for SpMM and SpGEMM and added heterograph support, demonstrating deep competence in production-grade GPU acceleration. Based in the San Francisco Bay Area, she combines academic rigor with cloud-scale deployment experience, making her adept at moving novel GNN and GenAI research into practical, high-performance systems.
10 years of coding experience
11 years of employment as a software developer
Bachelor's degree Computer Science and Engineering, Bachelor's degree Computer Science and Engineering at University of DhakaDoctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at The Ohio State University
English, Bengali, Hindi
Github Skills (8)
cuda10
sparse-matrix10
sparse-array10
sparse-file10
graph-neural-network9
cprogramming-language9
c-language9
deep-learning9
Programming languages (4)
Github contributions (5)
dmlc/dgl
Apr 2021 - Sep 2022
Python package built to ease deep learning on graph, on top of existing DL frameworks.
Role in this project:
Back-end Developer / CUDA Engineer
Back-end Developer / CUDA EngineerContributions:128 reviews, 23 commits, 38 PRs in 1 year 5 months
Contributions summary:Israt primarily focused on adding CUDA support for sparse matrix operations within the DGL (Deep Graph Library) framework. Their contributions included implementing CUDA kernels for Sparse Matrix multiplication (SpMM), SpGEMM, and related operations like summation and masking. These efforts involved debugging and optimizing CUDA code, addressing issues related to data type support, and refactoring existing functionalities to leverage GPU acceleration, specifically utilizing cuSPARSE. The user also contributed to adding heterograph support within the CUDA kernels.
pytorchpythondeep-learningmachine-learninggraph-neural-networks
isratnisa/graphstorm
May 2023 - Oct 2023
Enterprise graph machine learning framework for billion-scale graphs for ML scientists and data scientists.
Contributions:105 pushes, 19 branches in 5 months
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.