M Siddiqui
Software Engineer at Microsoft
Bellevue, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolM Siddiqui is a software engineer with nine years of experience building high-performance systems at Microsoft, focused on accelerating ML workloads through compiler and runtime work. He currently contributes to the AI Compiler team, leveraging MLIR and LLVM to turn neural network graphs into efficient machine code and driving hardware-software co-design. Previously he led technical efforts in ONNX Runtime and ML.NET—improving training performance, implementing kernels (including CUDA), and enabling ORTModule-backed PyTorch execution—while acting as release manager and mentor. His open-source contributions include critical work on ONNX loss operators and performance fixes in widely used repos like onnxruntime and dotnet/machinelearning. Beyond shipped features, he has a track record of production reliability improvements (e.g., VM Scale Set success-rate gains) and multiple patents, reflecting a blend of research instincts and pragmatic engineering. Based in Bellevue, WA, he’s equally comfortable debugging low-level kernels and shaping cross-team adoption of ML infrastructure.
9 years of coding experience
10 years of employment as a software developer
MS, Computer Science, MS, Computer Science at The University of Texas at AustinBS, Computer Science Honors, BS, Computer Science Honors at Purdue University
Github Skills (24)
pytorch10
c-language10
back-end-development10
net10
machine-learning10
dotnet10
onnx10
kernel10
asp-net10
deep-learning10
tensorflow10
kernel-mode10
dotnet-core10
cuda10
csharp10
Programming languages (8)
Github contributions (5)
dotnet/machinelearning
May 2018 - Dec 2019
ML.NET is an open source and cross-platform machine learning framework for .NET.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:7 releases, 213 commits, 555 PRs in 1 year 7 months
Contributions summary:M's commits focused on enhancing the FastTree algorithm within the ML.NET framework. They implemented feature mapping for disk transpose and improved the robustness of Generalized Additive Models predictors. The changes involved code modifications in FastTree and GamTrainer. They also made contributions to testing, by introducing new learners for the FastTree classifier and fixing existing tests related to the algorithm.
machine-learning-platformdotnetml-netmachine-learningcsharp
microsoft/onnxruntime
Mar 2020 - Aug 2021
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:213 reviews, 172 commits, 146 PRs in 1 year 5 months
Contributions summary:M contributed to the ONNX Runtime by merging pull requests related to ONNX submodule upgrades and implementing new kernels. They implemented and tested forward and backward kernels for the SoftmaxCrossEntropyLoss operator, including CUDA implementations to prevent divide-by-zero errors. Further work involved integrating the SoftmaxCrossEntropyLoss function into the BERT_LOSS function. They also made memory-aware gradient building improvements.
runtimetrainingtensorflowai-frameworkaccelerator
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.