Tamar Levy
AI Engineer at Sisense
Tel-Aviv, Tel-Aviv District, Israel
Summary
👤
Senior🎓
Top SchoolTamar Levy is an AI Engineer with over 13 years of software development experience, currently focused on backend AI algorithm work and applied ML at Sisense. She combines deep systems expertise—from early IA architecture and CPU optimizations to recent AVX2 performance engineering on VP9 codecs—with practical Python microservice development for AI-driven anomaly detection in manufacturing. While completing a master’s in machine learning with a thesis on NLP for endangered language revitalization, she has researched tokenizers and compared transformer and linear-algebra approaches, bridging research and production. Her background in open-source security software at Intel and hands-on profiling/optimization gives her a rare blend of low-level performance skill and high-level ML system design. Known for delivering reliable, production-ready systems, she also writes clear design documents and CI tooling, and is fluent in English and Hebrew. An understated strength is her track record of squeezing measurable gains from architecture and compiler-level fixes that directly improve real-world application performance.
12 years of coding experience
13 years of employment as a software developer
Master's degree to be completed in SEP 2022 Machine Learning, Master's degree to be completed in SEP 2022 Machine Learning at Tel Aviv UniversityB.Sc. Computer Science, B.Sc. Computer Science at Bar-Ilan University
English, Hebrew
Github Skills (9)
c1710
avx10
video-encoding10
c1110
performance-optimization10
assembler9
x869
assembly9
sse8
Programming languages (2)
Github contributions (5)
webmproject/libvpx
Oct 2013 - Apr 2015
Mirror only. Please do not send pull requests.
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:20 commits in 1 year 6 months
Contributions summary:Tamar's commits primarily focus on optimizing the VP9 video codec library, specifically targeting x86 architectures. Their work involves implementing AVX2 optimizations for various loop filter functions, Discrete Cosine Transform (DCT) calculations, variance computations, and Sum of Absolute Differences (SAD) algorithms. This includes processing 32 or 64 elements in parallel instead of 16, leading to significant function-level and overall user-level performance gains, especially for Atom processors. Additionally, the user addressed compiler bugs to ensure code correctness.
pull-requests
nlohmann/json
Sep 2018 - Sep 2018
JSON for Modern C++
Contributions:10 comments, 4 issues in 24 days
clangc-plus-plusmessagepackrfc-7159json-patch
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.