Kevin Chen
Software Engineer at NVIDIA
Sunnyvale, California, United States
Summary
🤩
RockstarKevin Chen is a software engineer with 7 years of experience specializing in deep learning inference and GPU-accelerated ML tooling, currently based in Sunnyvale and affiliated with NVIDIA. He has hands-on expertise integrating and extending TensorRT and ONNX ecosystems—contributing to ONNX Runtime, ONNX-TensorRT, and the TensorRT SDK by adding data-type support, new layer types, operator fixes, and test infrastructure improvements. His work spans both backend engineering and MLOps concerns, improving compatibility, build reliability, and production-ready operator support for model deployment. Kevin’s contributions to high-profile open-source projects demonstrate a pragmatic focus on performance, portability, and real-world deployment constraints across NVIDIA GPU platforms. An engineer who operates comfortably between low-level SDK APIs and higher-level runtime integration, he often surfaces subtle compatibility fixes that prevent production regressions.
7 years of coding experience
Github Skills (19)
c-language10
tensorrt10
machine-learning10
inference10
onnx10
nvidia10
deeplearning-ai10
deep-learning10
cprogramming-language10
api-design9
hardware-acceleration9
gpu-acceleration9
cuda9
algorithm7
algorithms7
Programming languages (5)
Github contributions (5)
onnx/onnx-tensorrt
Jan 2019 - Dec 2022
ONNX-TensorRT: TensorRT backend for ONNX
Role in this project:
Back-end Developer & MLOps Engineer
Back-end Developer & MLOps EngineerContributions:16 releases, 26 reviews, 238 commits in 3 years 11 months
Contributions summary:Kevin's primary contributions involve enhancing the ONNX-TensorRT backend. They implemented and fixed support for key operators, including those for image scaling, Gather, Unqueeze, Slice, and dynamic inputs. The user also focused on improving the software's compatibility, fixing compilation issues, and adding constraints for opset version requirements, indicating efforts to ensure the integration with and efficient usage of TensorRT. This suggests efforts towards deploying models via TensorRT.
nvidia-dockernvidiadeep-learningonnxbackend
NVIDIA/TensorRT
Sep 2019 - Sep 2022
NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.
Role in this project:ML Engineer
ML EngineerContributions:13 releases, 80 reviews, 66 commits in 3 years
Contributions summary:Kevin's commits focused on modifications to the `NvInfer.h` file, which is part of the NVIDIA TensorRT SDK. These changes included updates to the API, adding new layer types (like `kSLICE`, `kSHAPE`, `kPARAMETRIC_RELU`, and `kRESIZE`), and making adjustments to deprecated classes. This indicates the user was involved in enhancing and refining the core components of the TensorRT library, particularly in areas related to layer definitions and API functionality for deep learning inference.
deep-learning-inferencenvidiadeep-learningruntimeinference
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.