If I were Julius Caesar, I'd speak of myself in the 3rd person - as is in vogue for this section. While I have many astounding achievements, Caesar I definitely am not, so I'll talk in the 1st person. God knows my mother wishes I ate more humble pie, so here's looking at you, Mom! 😉 lol
I'm a Technology Consultant in a 30 year applied love affair with technology. As a Computer Engineer, I studied both hardware and software, graduating Cum Laude from Ain Shams University - Egypt's equivalent of MIT.
Even before graduation, I started working with NCR. I've since provided problem-solving - both strategic and tactical - to companies like Sony, Ericsson, Silicon Graphics (SGI), Wacom, & IBM. On the other end of the spectrum, I've worked with small business & startups, all via my boutique consulting practice, Larynx.
As a card-carrying Mensa genius with a documented 154 IQ, I throw myself wholeheartedly at whatever is the endeavor du jour - even though it may stray significantly from my education & prior experience. I LIVE for the challenge!
This is how I found myself at one point designing home theaters with 4K UHD screens & 11.2 channel Dolby Surround Sound.
My love of art is how I found myself the Marketing Director of an art gallery at another point, responsible for an $80 million inventory. That role had phenomenal fusion, creating a powerful trident that harnessed the synergy of my passion for art, technology, & marketing.
Over the last 6 years, Artificial Intelligence work has been taking more and more of my time: Machine Learning, with a particular focus on & passion for Neural Networks. Yes, I can roll up my sleeves & write Python code.
I can go on and on, but this is supposed to be an overview. So if you have a problem that keeps you awake at night with AI, know that I've solved quite a few. Heck, even if you think it's a long shot that I've solved something like this before, still try me. It's keeping you awake anyway, so what have you got to lose? 😏
- Project Management
Envisioned & supervised the creation of an inventory system with laser barcoded SKUs for an $80 million art collection - the largest of its kind in the USA.
As a Solutions Architect/Systems Integrator, I lead the team that integrated SGI (Silicon Graphics Inc) workstations, an industrial laser scanner, a Wacom tablet, & specialized German textile design software for Egypt’s premier textile company. The solution was the first of its kind in Egypt, the Middle East, & Africa.
- MEAN stack ( Mongo, Express.js, Angular.js/React.js, Node.js)
- Front-end : Javascript, jQuery, HTML5, CSS3, Bootstrap,Vue/Vuex
- Database : MySQL/MS SQL/Postgresql Administration, SQL optimization
- PHP Developer(Laravel Framework, CodeIgniter, Wordpress)
- Django, Web Scraping
- Ruby on Rails
- Python
My go to language for coding Neural Networks (NN), using Keras with a TensorFlow backend. Well-versed in using NumPy, SciKit (sklearn) & Pandas to preprocess data to clean, normalize & feature scale before feeding it to the Machine Learning model.
Adept at utilizing MatPlotLib and PyPlot for visualization.
Built a NN that predicted bank customer churn based on the the customer’s sex, geolocation, annual income, number of active accounts, & how long they had been with the bank.
Constructed a Convolutional Neural Network (CNN) for image classification. Added dropout layers to reduce overfitting, boosting accuracy from 82% to 92%.
Created a Recurrent Neural Network (RNN) to predict stock prices using Long Short Term Memory (LSTM) layers.
Crafted a hybrid model for fraud detection on credit card applications: a Self Organizing Map (SOM), followed by a NN.
Created a movie recommender system by implementing a Restricted Boltzmann Machine (RBM) using PyTorch. In this unsupervised learning model, Contrastive Divergence and Markov Chain Monte Carlo (MCMC) methods were used for predictions.
Took the movie recommender to the next level by restructuring it as an AutoEncoder. This unsupervised learning model can predict how a viewer will rate a movie with an error within one star in a 5 star system. Currently experimenting with the level of stacking, i.e. how many levels deep the NN is, as well as the number of neurons in each level, to further improve accuracy.
OKTA
