Li He

- Manage 3 teams (15+ people) to research and develop machine learning, image processing and computer vision algorithms for 4 eBeam products. - Team is 70% PhD and based on California, Michigan, and India - Applications include object detection, classification, segmentation, style transfer, registration, super resolution, denoise, tone mapping, NLP etc. - Coordinate vertical R&D (software, hardware, marketing, application) and horizontal R&D (other divisions, central AI team, contractors, professors)

- Manage a 10 people team to develop machine learning algorithms for 4 eBeam products - Lead 15 people team to develop algorithms of company’s first eBeam inspection product, including machine learning train/inference infrastructure and image process pipeline - Help KLA to grow from 0% EBI market share – the multi-million product is already shipped to 7 major customers - Coordinate cross function R&D (software, hardware, marketing, application) - Research: 3 US patents and multiple papers

- Led 5 people team to design algorithms of company's first deep learning platform - The deep learning engine is used on multiple products, priced at $2~$30 million, including ebeam, bright field / dark field inspection. - Applications include object detection, image classification, segmentation, image style transfer, registration, super resolution, image quality etc. - Research: 2 US patents and multiple papers

- Developed company's first machine learning engine, Intelligent Classification Engine (ICE) - ICE is associated with 3 US patents and 1 paper, has improved defect classification on multiple products, including ebeam, bright field / dark field and recital inspection. - Developed new image processing algorithms such as multi-layer image segmentation algorithm, image quality metric and feature extraction algorithm.

- Study centered on using statistical machine learning to improve patch based image super-resolution, improved over the state-of-the-art results (CVPR 13 paper) - Designed a neutron image denoising and restoration algorithm for Oak Ridge National Lab. - Developed a color image segmentation and emotion transfer method and evaluated the result based on user study. - Investigated sparse coding based linear unmixing algorithms and its application to anomaly detection of hyperspectral images and event detection in smart power system grid. - TA Courses: Computer System Fundamentals, Digital Image Processing and Pattern Recognition.

- Developed DigiWand™ - a wireless tablet pen gesture recognition software for Thinkpad laptop.

- Co-developer of Cy-Net™ wireless sensor network stack. - Product development – coordinate the development of an automatic gas meter reading product, helped the company to obtain the first 100K MCU production order in China.

- Designed a fast image analysis algorithm to recognize 2D bar code using a low configuration hardware platform.