Nour Islam Mokhtari

I am the general manager of PYCAD. We build web based medical imaging platforms with custom DICOM viewers. We also integrate custom AI components in these platforms. You can check out some of our projects on our portfolio page: https://pycad.co/portfolio/ Interested in working with us? DM me!

At Orpalis I worked on building next generation OCR (Optical Character Recognition) systems. My work consisted of applying deep learning techniques to recognize text in images of different formats (Cheques, serial numbers, ...). It also consisted of doing constant research about state of the art in text detection and recognition in images. Technologies used : Python, C++, C#, Tensorflow, Pytorch, ONNX.

In this role I worked on developing a business intelligence tool using Python and Dash. The goal of the application is to accelerate drug comparaison and cost prediction. The tools that I used during this time are : Python, Dash, AWS and DynamoDB.

As a machine learning consultant I help Alltech's clients plan and execute on machine learning based projects.

Tandemlaunch is an entrepreneurship program for aspiring entrepreneurs who want to build their own company.

My work is focused on developing AI solutions for industrial applications. So far, I have worked on the following tasks : 1/ Developing an image classification toolbox that uses various tools to classify images (deep learning, SVM,...). This toolbox is programmed using C++ and some libraries such as OpenCV. 2/ Developing a pipeline for object detection using Tensorflow object detection API. I have put in place a pipeline so that we can use pretrained models from the aforementioned API to detection objects in images coming from the industry. Some of the deep learning models that we used so far include : SSD (Single Shot Multibox Detector) and Faster-RCNN. We use this pipeline for 3 different main tasks : preparing data, training and then inference in production. For the first 2 tasks, we use Python and for inference we use tensorflow C++ API after it was built from source. 3/ Developing an OCR (optical character recognition) system that uses deep learning at its core. The system is divided into 2 main parts : text detection and text recognition. We used deep learning models for both parts. 4/ Using some data visualization tools in Python to visualize and make sense of data.

As an engineering intern in Diota company, I am working on different problems related to classification and recognition for industrial visual inspection. My work includes : 1) Developing machine learning solutions to classify different cases in mechanical parts (ex : there is a screw/there is no screw) 2) Developing a pipeline for visual inspection of any type of defects/cases, where a technician could send an image as input to the pipeline and the output would be the class corresponding to that specific case. I am trying to find a generic framework that could be applied to many cases and not just a specific one.

I did an internship at this company where I implemented an image labeling tool in C++ and then used it to label defects in images of wooden boards. I also used a small dataset of these images to train a convolutional neural network using Keras and Tensorflow.

During my internship at this company, I worked closely with control systems engineers. My main focus was instrumentation and control systems in the field of natural gas exploitation.

During my internship at this company, I worked on different control systems and I explored the instrumentation used in the brick manufacturing industry.