Ming Lin
Ming Lin

Senior Principal Applied Scientist

I lead applied research and engineering across robotics, reinforcement learning, agentic AI, and efficient AI systems. At Oracle, I helped establish the Oracle Cloud Robotics Laboratory and serve as its founding scientist and Seattle site lead.

My work connects advanced machine learning with reliable deployment—from large-scale model training and inference to embodied intelligence.

Projects →    Publications →    Experience →

Selected Publications

DRTriton: Large-Scale Synthetic Data Driven Reinforcement Learning for Triton Kernel Generation. NeurIPS 2026. arXiv · Code

DRPO: Efficient Reasoning via Decoupled Reward Policy Optimization. ICLR 2026. arXiv · Code

Beyond Gaussian Pyramid: Multi-skip Feature Stacking for Action Recognition. CVPR 2015. Article · arXiv

Feature Interaction Augmented Sparse Learning for Fast Kinect Motion Detection. IEEE TIP 2017. Article · BibTeX

Zen-NAS: A Zero-Shot NAS for High-Performance Image Recognition. ICCV 2021. Article · arXiv

DeepMAD: Mathematical Architecture Design for Deep Convolutional Neural Network. CVPR 2023. Article · arXiv · GitHub

Browse the complete publication list →