Renchu Wang
Data Scientist At GenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRenchu Wang is a data scientist and systems-minded ML engineer with 7 years of experience, currently leading multimodal document parsing and source-code translation efforts on C3 AI’s GenAI team. A recent Georgia Tech MS graduate with a BS in Electrical Engineering from NTU, he blends research-grade innovations (transformer masking and scalable quantum compilers) with production engineering—reimplementing a multimodal parser used by 85% of internal data scientists and improving its throughput 20x. He’s the primary contributor to the koila library that implements lazy tensors to mitigate PyTorch OOM issues, reflecting a practical focus on memory-efficient ML systems. Past roles span reinforcement-learning solutions at MediaTek, scalable pandas backends at Ponder, and automation for FPGA/green-compliance pipelines, showing strength in cross-domain problem solving and performance optimization. Colleagues describe him as someone who prefers to reimplement and truly understand systems from first principles: "What I cannot create, I do not understand."
7 years of coding experience
3 years of employment as a software developer
Bachelor of Science - BS, Electrical Engineering, Bachelor of Science - BS, Electrical Engineering at National Taiwan UniversityMaster of Science - MS, Computer Science and Engineering, Master of Science - MS, Computer Science and Engineering at Georgia Institute of TechnologyHigh School Diploma, High School Diploma at Taipei Municipal Jianguo High School
Github Skills (6)
memory-management10
pytorch10
machine-learning10
deep-learning10
python10
lazy-evaluation10
Programming languages (15)
Github contributions (5)
rentruewang/aioway
Nov 2021 - Jan 2023
AI on the way. An auto deep learning pipe dream. An RDBMS approach to deep learning. Declarative, explainable, scalable, optimizable, easy to deploy, all that good stuff.
Role in this project:
ML Engineer
ML EngineerContributions:79 commits, 69 PRs, 612 pushes in 1 year 2 months
Contributions summary:Renchu primarily contributed to the development of the `koila` library, which aims to address PyTorch's out-of-memory errors. Their work involved implementing core functionalities of the library, including the `LazyTensor` class, arithmetic operations, and various PyTorch methods. The contributions focused on creating a system for lazy evaluation and memory management within the PyTorch framework, as demonstrated by the addition of `DelayedTensor` and a refactor of existing interfaces.
aideep-learningrdbmspytorchlazy-evaluation
rentruewang/brief
Jan 2019 - Apr 2021
Text summarization with PyTorch
Contributions:22 commits, 29 pushes, 3 branches in 2 years 3 months
pytorchtext-summarizationpythonpaper-implementationsreinforcement-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.