Duong Nguyen
Senior AI Engineer at Samsung SDS
Hà Nội, Vietnam
Summary
🤩
Rockstar🎓
Top SchoolDuong Nguyen is a Senior AI Engineer with a PhD in Computer Science from UC San Diego and seven years of hands-on experience building production AI systems and ML tooling. Based in Hà Nội, he develops enterprise-grade solutions at Samsung SDS for document understanding, RAG chatbots, and LLM benchmarking while contributing upstream to Hugging Face projects such as diffusers and transformers—notably enhancing Flax-based Stable Diffusion, DreamBooth, and resource-efficient training scripts. His background blends academic research with practical systems work from roles at eBay and Facebook, where he automated CI/ML deployment and indexed codebases to improve service traceability. Earlier work in advertising applied topic modeling and optimization to triple CTR, reflecting a strong track record of measurable impact. Colleagues describe him as “not a Bayesian,” hinting at a pragmatic, experimental approach to model design and evaluation.
7 years of coding experience
8 years of employment as a software developer
University of California, San Diego
Github Skills (23)
transformers10
pytorch10
python10
pre-trained-model10
diffusion-models10
machine-learning10
stable-diffusion10
bart10
diffusers10
diffusion-probabilistic-models10
flax10
text2image10
diffusion-probabilistic10
deeplearning-ai10
deep-learning10
Programming languages (3)
Github contributions (5)
huggingface/transformers
Jul 2022 - Oct 2022
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:
ML Engineer
ML EngineerContributions:15 reviews, 6 commits, 9 PRs in 3 months
Contributions summary:Duong primarily worked on improving and maintaining Flax-based machine learning examples within the Hugging Face Transformers repository. Their contributions involved fixing resource exhaustion errors in Flax example scripts, particularly when dealing with large datasets, and ensuring the efficient use of JAX/Flax for model training. They also generalized the decay_mask_fn for more flexible application and added a pretraining script for the BART model. Furthermore, the user addressed output issues in various Flax-based models.
pythonbertspeech-recognitionstate-of-the-artflax
huggingface/diffusers
Oct 2022 - Nov 2022
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX.
Role in this project:ML Engineer
ML EngineerContributions:6 reviews, 10 commits, 12 PRs in 13 days
Contributions summary:Duong contributed significantly to the implementation and refinement of diffusion models, particularly focusing on the Flax framework for Stable Diffusion. They added features like textual inversion and DreamBooth, demonstrating expertise in fine-tuning and adapting diffusion models. Their work also included adding examples and addressing issues with data types and sample batch sizes in these Flax-based training scripts.
pytorchartdeep-learningimage2imagestate-of-the-art
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.