Elen Kalda
Software Engineer at Arm
Cambridge, England, United Kingdom
Summary
🤩
RockstarElen Kalda is a software engineer based in Cambridge with 10 years of experience focused on low-level and performance-sensitive systems. Currently at Arm, she blends practical engineering with a sense of humor—her GitHub bio jokes aside, she contributes seriously to open-source compiler work. Notably, she has added and tested quantized multiply support in the Relay compiler for the widely used TVM deep learning stack, helping improve performance for quantized tensors on diverse accelerators. Elen excels at writing robust tests and integrating nuanced numeric behavior like saturation handling, making her work valuable for production ML deployments. Her background suggests a strong aptitude for compiler internals and performance optimization across CPU, GPU, and specialized hardware.
10 years of coding experience
Github Skills (12)
compiler10
machine-learning10
deeplearning-ai10
compiler-compiler10
deep-learning10
tensor10
python10
numpy9
relay9
relayer9
quantization8
unit-testing8
Programming languages (4)
Github contributions (5)
apache/tvm
Oct 2019 - Jan 2023
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
ML Engineer
ML EngineerContributions:357 reviews, 29 commits, 85 PRs in 3 years 2 months
Contributions summary:Elen primarily contributed to the development and testing of quantized multiply operations within the Relay compiler. Their work focused on supporting quantized tensors by implementing a multiply operator and incorporating changes from related pull requests. The user's contributions involved writing and testing code to verify the functionality of quantized multiplication, including tests for different input/output parameters and saturation handling. They added support for a quantized multiplication to the Relay operator, specifically focusing on performance optimization within the deep learning compiler stack.
metalvulkancompilertensoropencl
ekalda/tvm
Oct 2019 - Apr 2024
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Contributions:75 pushes, 69 branches in 4 years 6 months
cpugpu-programmingcudagpu-accelerationtvm
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.