Dheeraj Peri
Senior Deep Learning Software Engineer at NVIDIA
Sunnyvale, California, United States
Summary
🤩
Rockstar🎓
Top SchoolDheeraj Peri is a Senior Deep Learning Software Engineer based in Sunnyvale with nine years of experience building and optimizing production-grade computer vision and reinforcement learning systems. At NVIDIA he has shipped performance-focused work—optimizing SSD object detectors, designing TensorRT plugins, and adding quantization-aware training support to flagship DeepLearningExamples—bridging research and engineering for real-world deployment. His open-source contributions include extending tensorflow-onnx with per-channel quantization and half-pixel resize support and adding TensorRT conversions in PyTorch tooling, reflecting deep expertise in model conversion and inference optimization. Comfortable across Python, C++, CUDA, TensorFlow and PyTorch, he pairs academic research in multi-modal learning from RIT with hands-on product experience from earlier roles. Notably, he combines low-level backend work (custom TensorRT layers, power-layer conversions) with a strong interest in generative and adversarial methods and deep reinforcement learning, bringing both systems-level rigor and algorithmic curiosity to ML engineering.
9 years of coding experience
3 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and ScienceMaster's degree Computer Engineering, Master's degree Computer Engineering at Rochester Institute of TechnologyDeep Reinforcement Learning Nanodegree Computer Science, Deep Reinforcement Learning Nanodegree Computer Science at Udacity
English, Telugu, Hindi
Stackoverflow
245reputation
56kreached
2answers
4questions
Github Skills (22)
pytorch10
c-language10
python10
convert10
converting10
tensorrt10
machine-learning10
onnx10
deep-learning10
tensorflow10
quantization10
computer-vision10
cprogramming-language10
model-optimization9
keras9
Programming languages (3)
Github contributions (5)
pytorch/TensorRT
Oct 2020 - Jan 2023
PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT
Role in this project:
Back-end Developer
Back-end DeveloperContributions:873 reviews, 675 commits, 501 PRs in 2 years 4 months
Contributions summary:Dheeraj implemented support for power layer conversion and other minor edits related to READMEs. They added conversion support for `torch.narrow()` and implemented test cases for element-wise operations. The user added support for 3d convolution and provided several testcases in this context.
cudapytorchjetsonnvidiadeep-learning
onnx/tensorflow-onnx
May 2020 - Feb 2022
Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX
Role in this project:ML Engineer
ML EngineerContributions:3 reviews, 6 commits, 6 PRs in 1 year 9 months
Contributions summary:Dheeraj primarily focused on modifying and extending the functionality of the `tensorflow-onnx` repository, particularly in the area of quantization and model conversion. Their contributions include adding support for half-pixel transformations in the resize operation and incorporating per-channel quantization within the QDQ (Quantize and Dequantize) framework. Furthermore, the user fixed issues and improved the codebase by modifying the existing test infrastructure and introducing features related to quantization aware training. The user also made some edits related to constant folding.
tensorflowjsexportdeep-learningonnxkeras-tensorflow
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.