Richard Shin
Research Scientist at Google DeepMind
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRichard Shin is a research scientist and PhD candidate from UC Berkeley with 15 years of experience at the intersection of machine learning and programming languages, specializing in neural program synthesis. He has held research and engineering roles across industry leaders—Google, Microsoft, and DeepMind—applying large language models to conversational assistants and production ML systems like Tensor2Tensor and Ray. His work spans core research, production engineering, and infrastructure (CI/CD and build tooling), reflecting a rare ability to move ideas from papers to deployed systems. At Microsoft he helped prototype Business Chat and developed LLM-driven semantic parsing techniques; on open-source projects he improved model performance and usability through advanced convnets and augmentation. Based in Berkeley, he combines deep academic training with hands-on contributions to widely used ML toolkits, and his background includes early security and tooling work under Dawn Song that continues to inform his focus on robust, verifiable models.
16 years of coding experience
14 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at University of California, Berkeley
English, Korean
Github Skills (17)
python10
machine-learning10
build-system10
distributed-systems10
cicd10
mask-rcnn10
deep-learning10
tensorflow10
faster-rcnn10
ray10
devops10
dockers9
docker9
parallel-processing9
computer-vision9
Programming languages (10)
Github contributions (5)
ray-project/ray
Apr 2017 - May 2017
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Role in this project:
Back-end & DevOps Engineer
Back-end & DevOps EngineerContributions:4 commits, 5 PRs, 6 pushes in 1 month
Contributions summary:Richard made several contributions related to the core functionality and infrastructure of the Ray project. They addressed issues within the Ray runtime, including fixing an object serialization problem and modifying how the scheduler address is returned. They also implemented and maintained the CI/CD pipeline by setting up and checking clang-format in Travis CI. Furthermore, the user contributed to the build system by creating a script to build manylinux1 .whl files.
aimachine-learningraydistributedparallel
tensorflow/tensor2tensor
Jun 2017 - Jul 2017
Library of deep learning models and datasets designed to make deep learning more accessible and accelerate ML research.
Role in this project:ML Engineer
ML EngineerContributions:7 commits, 4 PRs in 14 days
Contributions summary:Richard primarily contributed to the implementation and improvement of machine learning models within the TensorFlow framework. Their work includes adding image augmentation techniques for CIFAR-10, incorporating advanced convolution methods like fused subseparable convolutions, and integrating shake-shake blocks for the Shake-Shake model. They also addressed bugs and improved error messages, demonstrating a focus on model performance and usability.
deep-learningmachine-learningmachine-translationreinforcement-learningtpu
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.