Pauline Sho
Senior Software Engineer at Google
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolPauline Sho is a Senior Software Engineer in San Francisco with four years of professional experience building firmware and ML inference tooling at companies including Google and Fitbit. She bridges low-level embedded systems and machine learning, contributing to TensorFlow projects to add INT4 tensor support and optimize TFLite Micro for resource-constrained targets—a sign she focuses on squeezing performance and memory out of constrained hardware. Her background in firmware and electrical engineering (Cornell, Magna Cum Laude) informs pragmatic design choices from silicon to production code, and she’s extended that through an online MS in Computer Science from Georgia Tech. At Google she moved from software engineering into a senior role while continuing hands-on changes to famous open-source projects, demonstrating both product impact and community-minded engineering. Notably, she implemented precision-reduction passes and efficient 4-bit unpacking loops, work that meaningfully improves model quantization for embedded ML.
4 years of coding experience
7 years of employment as a software developer
Online Master of Science - MS, Computer Science, Online Master of Science - MS, Computer Science at Georgia Institute of TechnologyBachelor's Degree (Magna Cum Laude), ENG - Electrical & Computer Engineering, Bachelor's Degree (Magna Cum Laude), ENG - Electrical & Computer Engineering at Cornell University
English, Chinese, Malay
Github Skills (21)
convolutional-neural-networks10
c-language10
python10
machine-learning10
ml10
micro10
deep-learning10
tensorflow-lite10
tensorflow10
deep-neural-networks10
neural-network10
quantization10
cprogramming-language10
sys9
embedded9
Programming languages (4)
Github contributions (5)
tensorflow/tflite-micro
Apr 2022 - Nov 2022
Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).
Role in this project:
ML Engineer
ML EngineerContributions:142 reviews, 21 commits, 34 PRs in 7 months
Contributions summary:Pauline primarily contributed to the development and improvement of the TensorFlow Lite Micro framework, focusing on convolutional neural network (CNN) operations. They addressed build errors related to Hexagon targets, adding support for int32 biases in conv16x8 operations and enabling support for the quantize(uint8<->int8) operations. They also refactored code related to registration and added the new GetOptionalTensorData API. The user also worked on implementing a Python interpreter for TFLM, including support for custom operators and correcting a destructor bug.
signalml-modelslow-powerprocessorsdeployment
tensorflow/tensorflow
Jul 2022 - Nov 2022
An Open Source Machine Learning Framework for Everyone
Role in this project:ML Engineer
ML EngineerContributions:7 commits, 2 comments in 4 months
Contributions summary:Pauline's commits primarily focus on enhancing the TensorFlow/TensorFlow repository to support INT4 tensor operations within various convolutional and fully connected layers. They implemented support for INT4 tensors in depthwise convolution and fully connected layers within TFLite/TFLM shared reference kernels. Furthermore, the contributions include optimizing unpacking loops for 4-bit data and adding a pass to reduce tensor type precision if applicable. These changes improve model quantization and potentially reduce the memory footprint.
pythondata-sciencedeep-learningmlmachine-learning
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.