Crefeda Rodrigues
Staff Software Engineer at Arm
Manchester, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolCrefeda Rodrigues is a Staff Software Engineer based in Manchester with a PhD in ML systems and six years of experience optimizing machine-learning performance on Arm server-class AArch64 architectures. At Arm she has driven low-level performance work—improving oneDNN and TensorFlow integrations, matmul heuristics, threadpool schedulers and scratchpad allocation—to squeeze real-world speedups for ML workloads and MLPerf benchmarks. Her background blends academic research in HPC and Earth system modelling with hands-on engineering, giving her a rare fluency in both algorithmic performance and production toolchains. An active open-source contributor, she has enhanced widely used projects like tensorflow and oneDNN and adapted toolchains in ARM-software/Tool-Solutions to support TensorFlow 2 and oneDNN on AArch64. Colleagues rely on her for architecture-aware optimizations that turn research ideas into deployable, high-throughput systems.
5 years of coding experience
6 years of employment as a software developer
Master’s Degree, Advance Computer Science : Artificial Intelligence, Distinction, Master’s Degree, Advance Computer Science : Artificial Intelligence, Distinction at The University of ManchesterBachelor's Degree, Electrical, Electronics and Communications Engineering, 9.02, Bachelor's Degree, Electrical, Electronics and Communications Engineering, 9.02 at Manipal Institute of Technology
Github Skills (26)
performance-monitor10
c-language10
performance-analytics10
edn10
python10
machine-learning10
performance-measurement10
performance-analysis10
tensorflow10
deep-neural-networks10
neural-network10
performance-tuning10
arm10
performance-monitoring10
cprogramming-language10
Programming languages (5)
Github contributions (5)
ARM-software/Tool-Solutions
Jun 2020 - Jan 2023
Tutorials & examples for Arm software development tools.
Role in this project:
ML Engineer
ML EngineerContributions:5 reviews, 40 commits, 80 PRs in 2 years 7 months
Contributions summary:Crefeda primarily focused on enhancing the TensorFlow support within the tool-solutions repository. Their work involved adding and adapting patches for TensorFlow 2 support, integrating oneDNN, and building MLPerf benchmarks. They also modified build scripts and configuration files for TensorFlow, specifically targeting AArch64 architecture. The user's contributions demonstrate a focus on optimizing and expanding the capabilities of the tool-solutions for machine learning workloads.
development-toolsarmv8armv7armhfarm
uxlfoundation/oneDNN
Sep 2021 - Oct 2022
oneAPI Deep Neural Network Library (oneDNN)
Role in this project:Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:36 reviews, 10 commits, 9 PRs in 1 year 1 month
Contributions summary:Crefeda primarily contributes to the performance and architecture of the oneDNN library, specifically focusing on the aarch64 architecture. Their work involves implementing and optimizing matrix multiplication (matmul) functionality, including broadcast support and integration with Arm Compute Library (ACL). They also added support for threadpool runtime and updated JIT'ed reorder to handle inconsistent padding for the source. Furthermore, the user added a pooling primitive using the Arm Compute Library.
bfloat16sse41avx512openmpopencl
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.