Ramiro Leal-cavazos - Software Engineer at Google DeepMind

Ramiro Leal-cavazos

Software Engineer at Google DeepMind

Mountain View, California, United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Ramiro Leal-cavazos is a Senior Software Engineer in Mountain View with six years of experience building ML compilers and high-performance systems, currently at Google DeepMind. He combines a rigorous Mathematics and Physics background with deep practical expertise in compiler backends, contributing significant PJRT integration and buffer/memory-layout work to IREE and core features to Torch-MLIR. His work spans production-grade C/C++ systems (telescope controller at Stanford) to Python/Cython numerical algorithms, showcasing a rare fluency across low-level systems and high-level ML tooling. Ramiro is skilled at bridging PyTorch, XLA, and MLIR ecosystems to enable efficient model execution and kernel reuse, including exploration of XNNPACK within IREE. He brings a track record of shipping robust, well-tested code and improving interoperability across major open-source ML compiler projects. Colleagues value his combination of theoretical curiosity and pragmatic engineering that turns complex computational ideas into reliable, high-performance implementations.

6 years of coding experience

3 years of employment as a software developer

Associate of Science - AS Mathematics, Associate of Science - AS Mathematics at South Texas College

Master of Science - MS Computer Science, Master of Science - MS Computer Science at Stanford University

English, Spanish

Github Skills (21)

pytorch10

c-language10

runtimes10

python10

machine-learning10

mlr10

compiler-compiler10

xla10

compiler10

irr10

cprogramming-language10

ml9

mle9

testing8

cuda6

Programming languages (6)

C++LLVMHTMLMLIRJupyter NotebookPython

Github contributions (5)

llvm/torch-mlir

Sep 2021 - Jan 2023

The Torch-MLIR project aims to provide first class support from the PyTorch ecosystem to the MLIR ecosystem.

Role in this project:

Back-end Developer & ML Engineer

Contributions:1501 reviews, 84 commits, 277 PRs in 1 year 4 months

Contributions summary:Ramiro contributed to the Torch-MLIR project, providing support for the PyTorch ecosystem and MLIR ecosystem. The user implemented several python packages related to PyTorch's FX and Lazy Tensor Core for converting PyTorch code into MLIR. The user also made bug fixes and added tests, demonstrating involvement across multiple aspects of the project's development.

pytorchmlirtorchcompilerecosystem

iree-org/iree

Jun 2022 - Jun 2022

A retargetable MLIR-based machine learning compiler and runtime toolkit.

Role in this project:

Back-end Developer

Contributions:5 reviews, 1 commit, 3 PRs in 1 day

Contributions summary:Ramiro implemented and enhanced the IREE project's PJRT (Python/JAX Runtime) integration. Their contributions focused on adding and improving PJRT API methods, including buffer-related functions (size, unsafe pointer, ready event), and optimizing the execution flow. Key improvements included integrating memory layout data and support for host-to-device transfers, addressing non-dense input tensor layouts, and supporting new data types. These changes collectively streamlined IREE's compatibility with PJRT and the XLA ecosystem.

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.

Request Free Trial