Daniel Garvey
Member Of Technical Staff at AMD
Austin, Texas, United States
Summary
🤩
Rockstar🎓
Top SchoolDaniel Garvey is an ML-focused compiler engineer with 8 years of experience building production-ready ML runtimes and model deployment pipelines, currently a Member of Technical Staff at AMD in Austin. He spent five years at Nod Labs advancing IREE and SHARK Studio integrations—work that includes PyTorch→IREE conversion, dequantization passes, and compiling BERT/TensorFlow models for high-performance inference. Comfortable bridging research and product needs, he combines low-level compiler passes and memory-allocation fixes with practical deployment tooling. Earlier roles in operations and practice growth show he brings pragmatic project ownership and process improvements to engineering teams. Passionate about IR and ML, he’s an active open-source contributor whose work touches the well-known IREE ecosystem.
8 years of coding experience
11 years of employment as a software developer
High School Diploma, High School Diploma at Seven Lakes High School
Github Skills (22)
pytorch10
python10
torch-hub10
machine-learning10
mlr10
compiler-compiler10
tensorflow10
compiler10
irr10
torch-net10
runtimes9
compilation9
compile9
cprogramming-language8
deep-learning8
Programming languages (7)
Github contributions (5)
nod-ai/SHARK-Studio
Jul 2022 - Jan 2023
SHARK Studio -- Web UI for SHARK+IREE High Performance Machine Learning Distribution
Role in this project:
ML Engineer
ML EngineerContributions:161 reviews, 33 commits, 191 PRs in 5 months
Contributions summary:Daniel primarily contributes to fine-tuning and integrating machine-learning models within the SHARK Studio environment. Their work includes adapting and compiling TensorFlow models, specifically Bert, using IREE (MLIR) for deployment. They demonstrate experience with model download and integration, testing of model performance, and configuration of the build and deployment process. They also make changes to the download utilities.
pytorchcudaamdheterogeneousdeep-learning
iree-org/iree
Sep 2022 - Sep 2022
A retargetable MLIR-based machine learning compiler and runtime toolkit.
Role in this project:ML Engineer
ML EngineerContributions:27 reviews, 1 commit, 29 PRs in 1 day
Contributions summary:Daniel primarily contributes to the IREE compiler, focusing on integrating and optimizing the Torch (PyTorch) to IREE conversion pipeline. Their work includes adding passes to decompose complex operations, integrating dequantization operations from torch-mlir, and addressing bug fixes related to memory allocation. They also bumped the torch-mlir version and added a pattern for fpowi within the polynomial approximation pass.
mlirspirvvulkantensorflowcompiler
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.