Hong-yu Chiu
Senior Deep Learning Engineer at Keras
New Taipei, Taiwan
Summary
👤
Senior🎓
Top SchoolHong-yu Chiu is a Senior Deep Learning Engineer with nine years of experience specializing in model compression, computer vision, and image processing, currently developing NVIDIA’s TAO toolkit. He is a prolific open-source contributor to Keras—authoring over 100 PRs, adding core image ops like affine_transform, introducing dynamic int8 quantization and float8 training, and helping bring Lion optimizer and QLoRA workflows into the ecosystem. Previously he built production AI systems at Garmin and led 3D sensing and depth-completion research at Himax, delivering measurable cost savings and deployable models. Hong-yu combines strong research rigor (MEng with a 4.19 GPA) and production engineering know‑how, bridging PyTorch toolkits and Keras high-level APIs. Based in New Taipei, Taiwan, he is also recognized as an AI Google Developer Expert for Keras, reflecting both community leadership and deep technical impact. An interesting detail: his cross-disciplinary background includes undergraduate studies in biology, which informs a practical, data-driven approach to imaging problems.
9 years of coding experience
1 year of employment as a software developer
Master of Engineering - MEng, Electrical and Electronics Engineering, GPA: 4.19/4.3, Master of Engineering - MEng, Electrical and Electronics Engineering, GPA: 4.19/4.3 at National Chiao Tung UniversityBachelor's degree, Biology/Biological Sciences, General, Bachelor's degree, Biology/Biological Sciences, General at National Yang Ming University
Github Skills (13)
keras10
pytorch10
machine-learning10
deep-learning10
tensorflow10
python10
image-processing10
optimization9
neural-network9
jax8
natural-language-processing7
image-generation7
nlp7
Programming languages (4)
Github contributions (5)
keras-team/keras
Jan 2023 - Mar 2025
Deep Learning for humans
Role in this project:
ML Engineer
ML EngineerContributions:162 reviews, 178 PRs, 367 comments in 2 years 2 months
Contributions summary:Hong-yu contributed significantly to the core functionality of the Keras deep learning library, with a focus on image-related operations and optimization. They implemented the `affine_transform` operation, which is a core component for image manipulation. The user also refactored code, improved documentation, and fixed bugs across the repository. Furthermore, they contributed by introducing the Lion optimizer and improving the quality and accuracy of existing tools and functions.
pythondata-sciencedeep-learningneural-networksmachine-learning
keras-team/keras-io
May 2024 - Dec 2024
Keras documentation, hosted live at keras.io
Role in this project:ML Engineer
ML EngineerContributions:15 reviews, 8 PRs, 21 comments in 7 months
Contributions summary:Hong-yu primarily contributes to the Keras documentation, including examples demonstrating float8 training and inference, and LoRA/QLoRA fine-tuning. They develop and implement examples related to Stable Diffusion 3, including text-to-image generation and latent space exploration techniques. Their contributions showcase a focus on advanced machine learning techniques and model implementation within the Keras ecosystem.
deep-learninghostedmachine-learningtensorflowkeras
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.