[Nomination] Koki Mitsunami #205
Description
Select one:
- I am nominating myself for the PyTorch Ambassador Program.
- I am nominating someone else to become a PyTorch Ambassador.
Please confirm that the nominee meets the following requirements:
- The nominee is 18 years of age or older.
- The nominee agrees to abide by the PyTorch Code of Conduct.
- The nominee agrees to comply with the Linux Foundation Antitrust Policy.
- The nominee meets at least one of the qualifications listed in the PyTorch Ambassador Program Requirements.
Nominee Name
Koki Mitsunami
Nominee Email
Nominee's GitHub or GitLab Handle
mitsunami
(Optional) Organization / Affiliation
Arm
City, State/Province, Country
Cambridge, UK
Your Name
No response
Your Email (Optional)
No response
How has the nominee contributed to PyTorch?
- An active contributor to PyTorch repositories (e.g., commits, PRs, discussions).
- A speaker at PyTorch events or workshops.
- A PyTorch user group organizer or meetup host.
- A researcher or educator using PyTorch in academic work or training.
- An active leader in the PyTorch community with at least one year of experience in:
- Organizing events (virtual/in-person).
- Speaking at AI/ML conferences.
- Mentoring others in PyTorch.
- Creating technical content (e.g., blogs, videos, tutorials).
🏆 How Would the Nominee Contribute as an Ambassador?
As a PyTorch Ambassador, I would bring a unique ability to bridge the gap between cutting-edge deep learning research and practical deployment for production, particularly in resource-constrained environments like mobile and embedded systems. With a strong track record of making advanced ML concepts accessible, I have consistently empowered developers to bring models from experimentation to real-world applications.
I have authored in-depth technical blogs on topics such as optimizing inference on Arm architectures, efficient deployment strategies for mobile platforms, and using ML-Agents for mobile gaming. Through online tutorials, I’ve also guided developers in deploying AI models to mobile, demonstrating their ability to translate ecosystem-agnostic deployment challenges into actionable workflows for the broader ML community.
As a PyTorch Ambassador, I plan to:
- Lead community education by creating content and tutorials that explain advanced workflows like post-training quantization, QAT, and PyTorch-to-ExecuTorch conversion.
- Mentor developers and researchers, especially those working on mobile AI, edge computing, and gaming applications, through office hours, code reviews, and Discord or forum engagement.
- Collaborate across the ecosystem, surfacing feedback to the PyTorch team and helping shape tools that improve deployment interoperability, performance, and ease of use.
- Deliver talks that provide hands-on guidance for getting PyTorch models ready for production on diverse hardware.
My cross-framework experience and community-minded approach make me an ideal ambassador to help PyTorch thrive across research, development, and deployment domains, especially where performance and portability are key.Any additional details you'd like to share?
I have been deeply engaged in educating and supporting the ML and developer community through blogs, public talks, and tutorials — especially in the areas of model deployment, hardware acceleration, and cross-framework integration for mobile and edge applications.
[Blogs & Tutorials]
- Blog on PyTorch app with Android NNAPI – In this blog, I walk through how to enable and utilize Android NNAPI with PyTorch to accelerate inference on mobile devices, covering integration tips and performance benchmarks.
https://community.arm.com/arm-community-blogs/b/ai-blog/posts/improve-pytorch-app-performance-with-android-nnapi-support-386430784
- Blog series on ML-Agents with Unity for Mobile Gaming – I’ve written blog posts exploring reinforcement learning and inference optimization for mobile game environments using Unity ML-Agents.
https://community.arm.com/arm-community-blogs/b/mobile-graphics-and-gaming-blog/posts/1-unity-ml-agents-arm-game-ai
https://community.arm.com/arm-community-blogs/b/ai-blog/posts/p1-multi-agent-reinforcement-learning
- Blog on Arm CPU architectures – I’ve authored technical content to explain Arm's new SIMD architecture for enabling increased AI capabilities.
https://community.arm.com/arm-community-blogs/b/architectures-and-processors-blog/posts/sve2
- Tutorials on a Android chatbot application – I’ve created online tutorials that walk through converting and deploying ONNX models to mobile, helping bridge the gap from PyTorch training pipelines to production-ready mobile apps.
https://learn.arm.com/learning-paths/mobile-graphics-and-gaming/build-android-chat-app-using-onnxruntime/
[Conference Talks]
- Optimizing Stable Diffusion for Mobile – I presented techniques for compressing and accelerating Stable Diffusion models for mobile inference, including quantization strategies at GDC (Game Developers Conference).
https://www.youtube.com/watch?v=1vnKPLFxs0g&list=PLKjl7IFAwc4SSrROtKwHtcidGv7F6Wwi-&index=9
- Talks at CEDEC (Japan), AI and Games Summer School (UK), and GDC (US) – I’ve given multiple talks on using ML-Agents for game development and deploying reinforcement learning models in interactive, real-time environments.
https://school.gameaibook.org/2023-school/
- Talk on Arm64EC, a new way of building apps for Windows on ARM – delivered a technical talk focused on modern compile workflows for Windows on ARM, which has direct relevance for deploying PyTorch models on ARM64 Windows devices.
https://developer.arm.com/Additional%20Resources/Video%20Tutorials/DevHub/Arm64EC%20-%20ABI%20for%20Mixing%20x64%20and%20Arm64
https://developer.arm.com/Additional%20Resources/Video%20Tutorials/DevHub/ARM64EC%20-%20A%20new%20way%20of%20building%20apps%20for%20Windows%2011%20on%20Arm
These efforts reflect my commitment to open knowledge-sharing, practical tooling, and supporting developers as they bring PyTorch models from research to production—especially in constrained or emerging platforms like mobile and embedded systems.