Thanalakshan Sivarasa
Software Engineer at University of California, Los Angeles
Colombo, Western Province, Sri Lanka
Summary
🤩
Rockstar🎓
Top SchoolThanalakshan Sivarasa is a software engineer based in Colombo, Sri Lanka with 10 years of engineering experience and a strong foundation in mathematics and electrical engineering from Jaffna Hindu College and the University of Moratuwa. He has recent hands-on machine learning and engineering internships at Yarl IT Hub, WindForce and Trade Promoters Limited, applying theoretical insight to production-focused problems. As an open-source contributor to the popular TensorLy project, he improved tensor decomposition tests and strengthened the mxnet backend, showing a focus on numerical robustness and QA. He brings a systems-thinking approach influenced by systems biology to backend and ML work, turning mathematical intuition into reliable code. Outside of engineering he’s a cyclist and self-described nerd, a detail that hints at his discipline and curiosity.
11 years of coding experience
1 year of employment as a software developer
Bachelor of Science - BSc, Electrical Engineering, Bachelor of Science - BSc, Electrical Engineering at University of MoratuwaMaths, Maths at Jaffna Hindu College
English, Arabic
Github Skills (10)
tensorrt10
tensor10
python10
decomposition10
testing10
backend9
machine-learning9
back-end9
mxnet8
pytest7
Programming languages (12)
Github contributions (5)
tensorly/tensorly
Jun 2020 - Jan 2023
TensorLy: Tensor Learning in Python.
Role in this project:
Back-end Developer & QA Engineer
Back-end Developer & QA EngineerContributions:99 reviews, 180 commits, 108 PRs in 2 years 7 months
Contributions summary:Thanalakshan primarily focused on updating and testing functions within the tensorly/decomposition library. They modified several test files, particularly related to the tucker and parafac decomposition methods. The user made adjustments to code related to the mxnet backend, including adding a test, and fixing a clip function. These changes show a focus on the robustness and correctness of the tensor decomposition methods.
tensor-factorizationtensor-decompositionpythontensor-learningtensor
meyer-lab/DE.jl
Jul 2020 - Dec 2020
The Julia solver for DE learning.
Contributions:70 PRs, 124 pushes, 73 branches in 5 months
solverjulia
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.