/ ABOUT
Yuhua (Bill) Chen.
I work on post-training for frontier models at Amazon AGI — reinforcement learning, long-horizon reasoning, and the training environments behind it. What I care about most is durable learning signal: gyms easy enough to bootstrap a model and hard enough to keep improving a strong one, with rewards you can actually trust. Before this, I led the AI behind autonomous MRI scanners at Q.bio. I'm drawn to problems where a new approach changes the underlying logic — not just nudges what already exists.
/ EXPERIENCE
2025 - PRESENT
Applied Scientist · Amazon AGI
Frontier model post-training: reinforcement learning, long-horizon reasoning, synthetic data, and gym infrastructure. Algorithm development for multi-step agentic capability.
2024 - APRIL 2025
Staff Machine Learning Engineer & AI Research Lead · Q.bio
Led development of multi-modal large language models (LLMs) and visual foundation models for healthcare applications. Spearheaded innovations in AI-driven medical imaging and established the AI Research Team. Collaborated with executives to shape the company's long-term AI strategy.
2023 - 2024
Senior Machine Learning Engineer · Q.bio
Engineered and deployed 3D AI models for fully autonomous MRI scanning systems, achieving sub-second inference times. Led projects that reduced scan times by up to 9x and improved image quality by 27x, directly enhancing product performance and attracting new investors.
2021 - 2023
Machine Learning Engineer · Q.bio
Developed and implemented AI models for MRI segmentation and fast image reconstruction, improving processing speeds by 3000x. Designed a scalable deep learning infrastructure for multi-GPU training, which became the cornerstone of the company's AI operations.
SUMMER 2020
Applied Research Intern, Medical Imaging · Nvidia Corp.
Pioneered a Network Architecture Search (NAS) method for 3D medical image segmentation, reducing search times from weeks to minutes. Conducted large-scale benchmarking on 28,000+ GPU hours, optimizing network performance for medical applications.
SUMMER 2017
Research Scientist Intern · VoxelCloud Inc.
Designed and implemented 3D neural networks for lung CT analysis, achieving state-of-the-art results. Contributed to prototyping and feature releases in collaboration with the engineering team.
SUMMER 2016
Research Scientist Intern · Philips Research North America
Developed neural networks to identify patients with congestive heart failure using time-series data from 140,000 patients, achieving an AUROC of 88.55. Applied AI techniques to real-world clinical data at scale.
2013
Research Intern · Samsung Advanced Institute of Technology
Led the development of the multi-atlas label fusion (MALF) segmentation algorithm for brain imaging, improving segmentation accuracy and advancing the state-of-the-art in medical imaging.
/ EDUCATION
2016 - 2021
University of California, Los Angeles (UCLA)
Doctorate of Philosophy (Ph.D.) in Bioengineering
2015 - 2016
University of Pennsylvania (UPenn)
Master of Computer & Information Technology
2011 - 2014
University of Pennsylvania (UPenn)
Master of Science in Bioengineering
2007 - 2011
Northeastern University (China)
Bachelor of Science in Biomedical Engineering
/ FOCUS AREAS