Tzach's Resume

• Lead the design and development of large scale features, ensure that the outcomes meets both requirements and quality.
• Design and development of a scalable multi-threaded engine for medical data visualization and interaction.
• Seamless integration with the existing UI systems.
• Performance optimization and offload work to the GPU.
• Create and maintain vast amount of unit tests.
• Re-write of existing SW components by using the engine.
• Tools and technologies: C++ 20, Boost, OpenCV, OpenGL 4.6 (core), GLSL Shaders, GoogleTest, C++/CLI.

• Develop a component that visualizes a CT-scan’s topographic which allows the user to adjust and select regions for a further NM-scan.
• Optimize memory consumption and fix ~40 [MB] of unmanaged memory leak per system’s scan’s flow.
• Migrate legacy image processing libraries to OpenCV.
• Investigate and solve issues from on-site/remote systems.

• Research and compare existing convolutional neural networks (CNN) for Facial Landmark detection.
• Train a Landmark detection model.
• Development of a C++ infrastructure on HiKey970 board, that supports loading models and running inferences on the GPU (MALI-G72).
• Integration into an existing embedded application.
• Performance evaluation – achieved x2 overall FPS improvement.
• Tools and technologies: C++ 14, CMake, ArmNN, MNN, OpenCV, Python, TensorFlow.
• OS: Linux.

• Research about datasets and models for Depth Estimation.
• Use transfer learning to train a Resnet50 model on a custom dataset (NYUv2 format) for human depth estimation.
• Develop an algorithm to convert the annotations from Bounding Box to Segmentation to eliminate redundant pixels in the calculation of the loss function.
• Analyze and improve the model accuracy by using data analysis tool (that developed by me) to visualize the data and compare different configurations and variations of the dataset and hyperparameters.
• Tools and technologies: Python, PyTorch, Numpy, OpenCV.

• Design and develop internal systems, for data annotation, visualization, and analysis.
• Tools and technologies:
• Frontend: React.js & Redux with Material-UI.
• Backend: C# Web API Core, Python, OpenCV, Numpy.

• Extend rendering techniques in a C++ Direct3D 9 engine to support rendering for 3D monitors.
• Development of a stereoscopic 3D rendering feature.
• Upgrade the existing engine’s infrastructure from Direct3D 9 to Direct3D 12 (Direct2D interop) from the design stage until implementation and integration.
• The infrastructure’s upgrade and code optimizations achieved x2 FPS improvement.
• Create automated tests to test the upgraded engine’s correctness.
• Tools and technologies: C++ 11, D3D9, D3D12, D2D11, OpenCV, IPP, Google-Test.

• Specialty: Artificial Intelligence, Digital Signal and Image Processing.
• B.Sc Courses: Cloud and Distribued Systems, Cryptography, Theory of Computation, Software Methodologies, Principles of Programming Languages, Assembly Language and System Programming, Operating System, Computer Network, Algorithmics, Data Structures, Digital Design, Numerical Analysis, Linear Algebra, Calculus, Discrete Mathematics, Logic, Ordinary Differential Equation.

Outstanding learning capabilities and exceptional skills.

Teach math and core software subjects for 1st year students that are dealing with learning disabilities.