Chenhao Li

Santa Clara, CA 95051 · (408) 585-8158 · lichenhao@shiroisakura.com

I am software engineer at Google Cloud AI. I have experiences both in machine learning and standard backend development. My interests lie in applying computer vision knowledge to our daily life. This includes but not limit to: smart cameras, self-driving car, and robotics. While implementing these applications, we always face another problem - data. Therefore, I also work on how to use limited amount of data and(or) biased to train a SOTA machine learning model.

Experience

Software Engineer

Google LLC, Sunnyvale
  • Improved the lightweight model accuracy by introducing the pseudo label generated by a high accuracy expensive teacher model on predicting speed sign.
  • Migrated data processing pipeline from simple Python multithreading to Apache Beam.
  • Created Google Cloud Vertex AI pipeline for machine learning training and inference purpose based on Kubeflow.
July 2020 - Present

Software Engineer Intern

Enjoyor Co. Ltd., China
  • Compiled Fully Convolutional One-Stage Object Detection algorithm.
  • Developed the algorithm with GIoU loss and a new mapping method from feature maps to input image.
August 2019 - September 2019

Lab Tutor

Donald Bren School of Information and Computer Sciences, UCI
  • Assisted over 50 students learning computer science in their homework.
  • Guided students how to hand-stimulate code and use debug mode to find their mistakes.
September 2018 - December 2018

Project

Grammatical Error Correction

  • Completed simple n-gram probability calculation on grammar validation over long paragraphs.
  • Implemented grammar error correction model for long sentences.
  • Tracked the performance of model with different embedding methods, e.g., naive one-hot embedding, BERT-based, GPT-based.
April 2021 - June 2021

Semantic Segmentation

  • Utilized PSPNet to do the semantic segmentation on a map of Chbar Mon, Kampong Speu Province of Cambodia.
  • Analyzed results over IoU and pixels wide accuracy.
April 2020 – June 2020

Histopathologic Cancer Detection

  • Implemented ResNet101 to predict 32x32 Histopathologic cancer image.
  • Achieved 95% testing accuracy.
September 2019 – December 2019

Optimization Method Validation

  • Reimplemented LookAhead optimization method in Tensorflow.
  • Validated the results with original paper.
September 2019 – December 2019

Handwritten Digit Recognition

  • Performed Decision Tree Classification, Random Forest Classification and Support Vector Machine on recognizing handwritten digits from Simion dataset.
  • Achieved nearly 98% accuracy on predicting the digits correctly.
January 2019-March 2019

Sentiment Analysis on Amazon Review

  • Implemented Convolutional Neural Network and Long-Short Term Memory to analyze the sentiment of Amazon reviews.
  • Achieved 83% accuracy.
January 2019-March 2019

Education

University of California, Irvine

Bachelor of Science, Bachelor of Arts
  • Computer Science - specialization in Intelligent Systems
  • Mathematics - specialization in Applied & Computational Mathematics
  • Quantitative Economics

GPA: 3.82

September 2016 - June 2020

Skills

Programming Languages & Tools

Interests

Apart from being a backend software engineer, I enjoy most of my time in playing video games. I am interested in both console and PC games and sometimes playing mobile games (Genshin Impact and Blue Archive are the two I am currently playing).

Feel free to reach out to me through ♿tt♿#1898 and my steam homepage .

And there are several games I really want you to try out. Some of them are very popular and some you may not like. But here is the list I recommend:

  • Xenoblade Chronicles series (switch)
  • Persona series (PS, PC)
  • Witcher 3 (All platforms)
  • Utawarerumono series (PS, PC)
  • Fire Emblem: ThreeHouses (switch)