Dec 2024 – Present · 1 yr 5 mos

• I oversee Deloitte’s global technical alliance with NVIDIA, engaging with senior executives and engineering leaders to co-create transformative solutions, influence product strategy, and scale delivery worldwide. By bridging strategy, innovation, and execution, I ensure that Deloitte and NVIDIA together accelerate enterprise transformation across industries. My leadership spans next-generation AI, large language models (LLMs), and high-performance computing, driving impact from silicon to service & product.

Jan 2024 – Present · 2 yrs 4 mos

• Head of the Deep Learning and Accelerated computing at Deloitte, I manage the NVIDIA alliance in Israel, leading client projects and training Deloitte colleagues in AI and accelerated computing globally. I Joined Deloitte with the acquisition of my company OpTeamizer, which I founded on 2015 and was an NVIDIA partner.

Aug 2023 – Present · 2 yrs 9 mos

• Chairman of the AI forum at the Israeli Hi-Tech Association

Jun 2023 – Present · 2 yrs 11 mos

• While Computer Vision has been a dominant force in the field of AI, NLP has recently started also to take center stage. The ability for machines to understand, interpret, generate, and interact using human language is becoming increasingly relevant in today's technologically-driven world. This is where our new course steps in.
• The course provides a comprehensive understanding of NLP using deep learning, right from the foundations of word vectors to the sophisticated applications of self-attention and transformers. The curriculum is meticulously designed to balance theory and hands-on experience, featuring practical tutorials on PyTorch and engaging insights into recurrent neural networks and natural language generation.
• In an era where AI is revolutionizing industries and redefining possibilities, mastering NLP is no longer a choice but a necessity. Looking forward to seeing you in class!

Nov 2021 – Present · 4 yrs 6 mos

Oct 2020 – Present · 5 yrs 7 mos

• Learn deep learning techniques for a range of computer vision tasks, including training and deploying neural networks. You will learn to:
• Implement common deep learning workflows such as Image Classification and Object Detection.
• Experiment with data, training parameters, network structure, and other strategies to increase performance and capability.
• Deploy your networks to start solving real-world problems.

Apr 2019 – Present · 7 yrs 1 mo

• NVIDIA DLI Certified Instructor for CUDA Python, teaching the course in Israel:
• This course explores how to use Numba—the just-in-time, type-specializing Python function compiler to accelerate Python programs to run on massively parallel NVIDIA GPUs.
• It teaches how to:
• Use Numba to compile CUDA kernels from NumPy universal functions (ufuncs).
• Use Numba to create and launch custom CUDA kernels.
• Apply key GPU memory management techniques.
• Upon completion, attendees will be able to use Numba to compile and launch CUDA kernels to accelerate Python applications on NVIDIA GPUs.

Oct 2018 – Present · 7 yrs 7 mos

Oct 2018 – Present · 7 yrs 7 mos

Oct 2018 – Oct 2020 · 2 yrs

Apr 2020 – Present · 6 yrs 1 mo

• NVIDIA DLI Instructor of the workshop for Deep Learning using Multiple Data Types.
• The workshop learning objectives:
• > Implement common deep learning workflows such as image segmentation and text generation
• > Compare and contrast data types, workflows, and frameworks
• > Combine deep learning-powered computer vision and natural language processing to start solving
• sophisticated real-world problems that require multiple input data types

Jan 2020 – May 2024 · 4 yrs 4 mos

• Professional Evaluator for the Israeli Innovation Authority, Office of the Chief Scientist (OCS).
• I perform evaluations of industry funding requests in the Computer Vision Software domain.
• I provide the OCS committee with information, comments and reports for the evaluation of the grant applications.

Oct 2019 – May 2024 · 4 yrs 7 mos

Mar 2019 – Present · 7 yrs 2 mos

• Lecturer of the Deep Learning course at the Ort Braude college of Software Engineering.
• The course is based on Stanford's excellent deep learning course.
• Course content:
• 1. Computer vision overview
• 2. Python/numpy tutorial
• 3. Image classification
• 4. Convolutional Neural Networks
• 5. Training Neural Networks
• 6. Deep Learning Hardware and Software
• 7. CNN Architectures (AlexNet, VGG, GoogLeNet, ResNet)
• 8. Recurrent Neural Networks
• 9. Practical Object Detection and Segmentation
• 10. Visualizing and Understanding
• 11. Video understanding

Nov 2018 – Oct 2019 · 11 mos

• Lecturer of the Digital Signal Processing course at ORT Braude, Software Engineering department

Feb 2017 – Oct 2018 · 1 yr 8 mos

Oct 2016 – Oct 2018 · 2 yrs

Oct 2015 – Oct 2018 · 3 yrs

• Lecturer of Android Development course

Sep 2015 – Present · 10 yrs 8 mos

• Lecturer of Heterogeneous Parallel Programming course.
• The course covers heterogeneous computing architectures, data-parallel programming models, techniques for memory bandwidth management, and parallel algorithm patterns.
• The programming languages of the course are OpenCL and CUDA.

Jul 2015 – Jan 2017 · 1 yr 6 mos

May 2015 – Oct 2015 · 5 mos

• BioSense Innovation Team

Mar 2014 – Jan 2017 · 2 yrs 10 mos

Mar 2013 – May 2024 · 11 yrs 2 mos

• Make it right. Then, Make it fast!
• Founder and CTO at OpTeamIzer Ltd.
• OpTeamizer provides the services of mentoring, consulting and implementation of projects where an expertise of accelerating applications performance is required, or where designing the architecture for such a system needs to be planned.
• Software optimization:
• CPU Optimizations using SIMD SSE/AVX, parallel programming.
• Usage of atomic commands, lock free data structures, etc.
• Finding hotspots, detecting and analyzing of various types of bottlenecks.
• GPU Development:
• Design, architect and implement new systems.
• GPU optimizations of OpenCL code, supporting high memory bandwidth requirements and high compute efficiency.
• Mentoring and tutoring development teams, teaching a 3 days course of hands on OpenCL.

Feb 2013 – Feb 2015 · 2 yrs

• Ranked each year as a Role Model
• Leading a scrum team in an agile environment, focused on reliability and performance of the Ultrasound system
• Surfacing, tackling and solving complex software bugs
• Designing mechanisms for preventing or early detecting software bugs according to the history of bugs found so far. Most challenging bugs were related to memory corruption.
• As a result of this effort, managed to achieve the lowest crash rate that the Ultrasound scanner ever achieved. ( More than x200 improvement)
• Defining the SW team workflow and introduction of Continuous Integration using Jenkins CI server
• Leading the automation of stress tests, logs analysis and crash dumps, all incorporated into Jenkins

Nov 2009 – Jan 2013 · 3 yrs 2 mos

• Ranked each year as a Role Model and was chosen to participate in a 2 years GE Excellence program.
• Architect and developer of GPU OpenCL image reconstruction, unique processing data transfer rates of 3GB/s.
• Parallel developer and designer in a large C++ codebase (1.5M lines of code), with more than 100 executing threads at runtime.
• Architect and developer of a highly efficient parallel tissue processing framework which maximizes CPU cache utilization
• Expert at analyzing hotspots/bottlenecks using Intel VTune
• SIMD Programming - Acceleration of C++ image processing algorithms by using assembly like vector instructions (SSE/AVX)
• Leading software projects with academic collaboration: CPU Profiler, C++ 11 mechanisms, Android voice recognition.
• Mentored the employees on a wide topics related to software engineering, efficient code, etc.
• Awarded for improving system lifetime by x10 by locating and fixing many memory leaks which caused memory fragmentation at the time the system was running on 32 bits.
• Awarded for solving a DICOM connectivity issue where the customers experienced a very long time required for sending exams.

Feb 2013 – Jun 2013 · 4 mos · Haifa

• Leading 2 students in a software project for General Electric.
• Project: C++11 advanced Mechanisms.
• Implementations of image processing pipeline, lock free data structures and the concept of hierarchical locking.

Jan 2013 – Jan 2018 · 5 yrs

Mar 2012 – Mar 2012 · 0 mo

• Lecturer in a 40 hours course, covering:
• CPU architecture
• Writing Efficient code
• SIMD programming (SSE)
• Profiling using VTune
• Parallel programming
• Unit testing
• Static Analysis

Feb 2012 – Jun 2012 · 4 mos · Haifa

• Leading 2 students in a software project for General Electric.
• Project: Android voice recognition.
• Incorporated Android voice recognition into a mobile android remote control application, used for controlling GE Ultrasound system.

Jan 2012 – Jun 2012 · 5 mos · Haifa

• Hired by the Technion, Civil Engineering department
• Improved by x100 a Flac3D software which models non-conventional wells. Without the optimization it would have been unusable due to long run time.

Feb 2011 – Jun 2011 · 4 mos · Haifa

• Leading 2 students in a software project for General Electric.
• Project: CPU profiler.
• Using the debugger API, sampled the callstack of the running process every few milliseconds and plotted its histogram for locating hotspots.

Jan 2011 – Jan 2011 · 0 mo

• Accelerated an existing CUDA implementation by a factor of x3

Jan 2007 – Jan 2009 · 2 yrs · Haifa

Jan 2007 – Jan 2009 · 2 yrs · Haifa

Jan 2006 – Jan 2007 · 1 yr · Haifa

Jan 2005 – Dec 2008 · 3 yrs 11 mos · Israel

• Investigated and predicted the scaling of mechanisms related to the CPU performance states and sleep states.
• Development using C#, Java, ASM, minimal Windows driver development and VBA(Excel) for data analysis.