Pengsheng Huang

Owner at beijing rongtang tech

Beijing, China
email-iconphone-icongithub-logolinkedin-logotwitter-logostackoverflow-logofacebook-logo
Join Prog.AI to see contacts
email-iconphone-icongithub-logolinkedin-logotwitter-logostackoverflow-logofacebook-logo
Join Prog.AI to see contacts

Summary

🤩
Rockstar
Pengsheng Huang is an experienced software engineer and entrepreneur with eight years in high-performance and machine-learning engineering, currently running Beijing Rongtang Tech and working as an SDE on Xiaomi's VLM team. He brings a background in HFT development and prior experience at SenseTime, combining low-latency systems thinking with practical machine learning deployment. Pengsheng contributes to prominent open-source ML tooling—most notably implementing Factor Transfer modules in the OpenMMLab mmrazor compression toolbox and adding CRD dataset support—showing fluency in model compression and research-to-code translation. Based in Beijing, he blends hands-on coding with product-minded ownership, often bridging research prototypes and production-ready pipelines. An engineer who moves between finance-grade performance constraints and cutting-edge vision models, he has a track record of shipping robust, reusable components in large ML ecosystems.
code8 years of coding experience
github-logo-circle

Github Skills (8)

model-compression10
pytorch10
distill10
python10
classification10
faster-rcnn9
mask-rcnn9
computer-vision9

Programming languages (2)

C++Python

Github contributions (5)

github-logo-circle
open-mmlab/mmrazor

Aug 2022 - Dec 2022

OpenMMLab Model Compression Toolbox and Benchmark.
Role in this project:
userML Engineer
Contributions:13 reviews, 7 commits, 14 PRs in 3 months
Contributions summary:Pengsheng primarily contributed to the implementation of Factor Transfer, a model compression technique, by adding the necessary modules and configurations. Their work involved defining the `Paraphraser` and `Translator` modules, which are core components of the Factor Transfer architecture, and integrating them into the `mmrazor` framework. The user also created configurations for pretraining and training Factor Transfer models within the context of CIFAR-10 and ResNet architectures. Furthermore, they introduced a Contrastive Representation Distillation (CRD) dataset wrapper.
pytorchsegmentationknowledge-distillationdetectionnas
fpshuang/mmrazor

Aug 2022 - Jun 2024

OpenMMLab Model Compression Toolbox and Benchmark.
Contributions:1 PR, 48 pushes, 18 branches in 1 year 10 months
pytorchtoolboxcompressionbenchmarkopenmmlab
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.
Request Free Trial
Pengsheng Huang - Owner at beijing rongtang tech