Niklas Gustafsson
Principal Program Manager PM Architect at Microsoft
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolNiklas Gustafsson is a Principal Program Manager and PM Architect at Microsoft with over three decades of engineering and leadership experience and a decade summarized in this profile focused on developer tooling and machine learning platforms. He helped shape Azure SDK strategy and contributed to cross-platform developer engagement, language runtime and concurrency innovations dating back to work that influenced async/await design. A hands-on engineer, Niklas contributes to prominent open-source projects such as TorchSharp and TensorFlow.NET (implementing low-level tensor storage, loss modules and Keras pooling/text preprocessing) and improved AutoRest code generation for Azure SDKs. He combines deep systems and compiler expertise with product-level operational improvements from startup to enterprise, and is noted for bridging research-level ideas into production developer tools. Based in Redmond, he brings a rare mix of standards‑level participation, large-scale platform strategy, and ongoing low-level ML library development.
10 years of coding experience
24 years of employment as a software developer
M.Sc. Computer Engineering, M.Sc. Computer Engineering at KTH Royal Institute of Technology
English, Swedish
Github Skills (25)
unit-testing10
operation10
package-development10
plugin-development10
tensorrt10
low-level10
memory-management10
machine-learning10
java10
javas10
keras10
code-generation10
tensorflow10
dotnet-core10
management10
Programming languages (13)
Github contributions (5)
dotnet/TorchSharp
Nov 2018 - Jan 2023
A .NET library that provides access to the library that powers PyTorch.
Role in this project:
Backend Developer
Backend DeveloperContributions:3 releases, 613 reviews, 1451 commits in 4 years 3 months
Contributions summary:Niklas primarily contributed to the core functionality of the TorchSharp project, with a focus on low-level tensor storage operations and the creation of modules for deep learning models, especially in the context of the PyTorch framework. They implemented and modified critical components related to tensor storage and module functionality. The user's work included the implementation of new modules for various loss functions, and ensuring the compatibility with Python tools.
pytorchnet-librarydotnetpythonpowers
SciSharp/TensorFlow.NET
Feb 2021 - Mar 2021
.NET Standard bindings for Google's TensorFlow for developing, training and deploying Machine Learning models in C# and F#.
Role in this project:ML Engineer
ML EngineerContributions:1 review, 17 commits, 9 PRs in 19 days
Contributions summary:Niklas contributed to the development of the TensorFlow.NET bindings, specifically focusing on Keras integration. They implemented and tested various pooling layers (GlobalAveragePooling1D, GlobalMaxPooling1D, GlobalAveragePooling2D, MaxPooling2D) within the Keras framework. Further contributions included adding text preprocessing APIs like Tokenizer and related unit tests, demonstrating a focus on preparing data for machine learning models.
deployingdotnetdevelopingscisharpdeep-learning
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.