Skip to content
View Kugman's full-sized avatar
🏆
Working hard
🏆
Working hard

Block or report Kugman

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Kugman/README.md

About Me

I am Efrat, a software engineer with deep experience in embedded systems, real time systems and C++. My work focuses on performance, reliability and low level problem solving. Over more than a decade in the industry I have designed and built systems that interact directly with hardware, optimize computation paths and handle strict timing requirements.

I enjoy challenges that demand precision thinking, from debugging complex race conditions to improving performance across different architectures. I care about clean design, predictable behavior and writing code that stands up to real world constraints. I am also interested in modern directions in computing, including GPU programming, parallel workloads and machine learning accelerators.

Outside of my professional work I spend a lot of time learning. I explore new technologies, study tools that catch my interest and build personal projects that help me grow as an engineer. I like experimenting, testing ideas and improving my technical understanding through hands on practice.

My GitHub is a place where I document this work. I share projects, try new approaches and keep track of what I learn. It reflects my curiosity, my practical mindset and my commitment to improving the way I build software.

Pinned Loading

  1. City-Traffic-Light-Synchronization-System City-Traffic-Light-Synchronization-System Public

    A GPU-accelerated traffic light management system using HIP (Heterogeneous-compute Interface for Portability) that optimizes traffic flow through intelligent synchronization and real-time adaptation.

    HIP 2

  2. GPU-Convolution-Engine GPU-Convolution-Engine Public

    Developed a GPU-accelerated 2D Convolution Engine in C++/CUDA, implementing core operations used in Computer Vision and Deep Learning. Optimized with shared memory and tiling, benchmarked against C…

    Cuda 2

  3. Real-Time-GPU-Log-File-Analyzer Real-Time-GPU-Log-File-Analyzer Public

    Tails a growing log file, and uses GPU to count error/warning patterns in real time.

    Cuda 3

  4. CudaMatrixEngine CudaMatrixEngine Public

    High performance GPU engine for matrix multiplication, a core operation in AI/ML. Implemented in c++ / CUDA, optimized memory access and computation for maximum throughput, achieving significant sp…

    Cuda 2

  5. Lock-free-Queue Lock-free-Queue Public

    High performance lock-free queue in C++ demonstrating real-time safe concurrency and non-blocking data structure for embedded and high performance systems.

    C++ 2

  6. LinkedIn-Games-Helper LinkedIn-Games-Helper Public

    🛠️ Chrome Extension for automatic solving of daily LinkedIn games! Includes Queens 👑, Tango ☀️, Zip 🐍, and Mini Sudoku 🔢. The code leverages advanced Backtracking algorithms and DOM Manipulation fo…

    JavaScript 5