Berkin Ilbeyi
Staff Software Engineer at Google
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolBerkin Ilbeyi is a Senior Staff Software Engineer at Google with 11 years of experience specializing in the intersection of computer architecture and JIT compilers for dynamic languages, currently pursuing a PhD at Cornell. He has a deep track record in compiler back-end engineering, notably improving memory space assignment and prefetching for XLA within TensorFlow and OpenXLA to boost ML model performance on accelerators. At Google he progressed from software engineer to senior staff, focusing on TPU/XLA compilation and low-level memory optimizations that reduce footprint and improve throughput. His work blends rigorous academic research—mentored by leading architecture labs—with production-scale systems engineering, reflecting internships at Oracle (Graal) and Google Platforms. Based in New York, he pairs systems-level thinking with hands-on contributions to widely used open-source ML compilers, demonstrating an uncommon fluency across hardware-aware compiler optimizations and runtime JIT techniques.
11 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD), Electrical and Computer Engineering, Doctor of Philosophy (PhD), Electrical and Computer Engineering at Cornell UniversityRobert CollegeBS, Electrical and Computer Engineering, Computer Science, BS, Electrical and Computer Engineering, Computer Science at Lafayette College
Turkish
Github Skills (14)
data-structures10
memory-management10
c-language10
tensorflow10
cprogramming-language10
performance-optimization10
python10
data-structure10
optimization10
machine-learning9
algorithm9
algorithms9
deep-learning8
multithreading7
Programming languages (2)
Github contributions (5)
openxla/xla
Aug 2019 - Jan 2023
A machine learning compiler for GPUs, CPUs, and ML accelerators
Role in this project:
Back-end Developer
Back-end DeveloperContributions:3 reviews, 224 commits, 1 comment in 3 years 6 months
Contributions summary:Berkin's commits primarily focused on implementing and improving the memory space assignment logic for the XLA compiler, a machine learning compiler. They made significant changes to the handling of bitcasts in memory space assignment, added support for limiting outstanding async copies, and refactored the prefetch interval picking process. Their contributions involved both refactoring existing code and introducing new features to enhance the performance of memory management.
compilercommunity-drivenmachine-learningmodular
tensorflow/tensorflow
Aug 2019 - Jan 2023
An Open Source Machine Learning Framework for Everyone
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:225 commits in 3 years 6 months
Contributions summary:Berkin primarily contributed to the XLA compiler within the TensorFlow project, focusing on memory space assignment and related optimizations. They implemented and refined algorithms for buffer allocation, cross-program prefetching, and memory-bound loop optimization. Significant work included enhancing the memory space assignment to improve performance and reduce memory usage in a variety of scenarios within the TensorFlow ecosystem. The commits also suggest an understanding of the interplay between memory management and the performance of machine learning models.
pythondata-sciencedeep-learningmlmachine-learning
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.