Marawan Shalaby

- Developed models for predicting purchase probability and optimal allocations for millions of customers to balance revenues and costs. - Worked extensively on the productionalization of predictive models by going through the full model development lifecycle including: - Idea Inception and Modelling - Implementation of REST APIs and connecting to OLTP and analytical databases - Deployment to Openshift and Kubeflow - Implementing CI/CD pipelines

Implemented projects for a wide variety of customers. Worked on both the data science and engineering sides. Projects include: - Fashion recommendation using triplet networks+ productionizing using docker and Rest API - Image segmentation using modified UNet networks. Productionizing using Azure ML Service - Chatbots using Microsoft LUIS + Kubernetes. Used Typescript and java for the backend - News aggregation using Elasticsearch and python

- Creating Deep learning recommender systems for recommending similar products to the ones the customer has interacted with - Creating an offer classification system for classifying products based on both images and text. - Technologies used: Python, Keras, tensorflow

- Working on Data Analysis with Spark/Hadoop to gain insights into hotel and review data - Web app development with Python, Twisted/Tornado, Mongo, PostgreSQL and jQuery

Design and Implementation of enterprise business applications using the IBM Business Process Management (BPM) stack with Master Data Management (MDM) backend. Working with the IBM BigInsights Big Data platform and backend integration with Salesforce backend. Languages used were Java and JS.

Development of Frontend and Backend for systems used in Digital Media Campaigns using Adobe AIR for frontend and Java and C# for the backend. Projects: IMAP: Electronic Wayfinding solution with web-based CMS, with advanced alogrithms for calculating multi-level shortest paths and displaying animated path on interactive multi-touch interface. Responsibilities: - System Analysis and Design - Frontend development using Adobe AIR - Backend development using Java (EJB3, JSF, Hibernate), C#

Frontend and backend web development using Java, Google Web Toolkit (GWT) ajax framework, Hibernate and Spring. ProjectsExample: Google Insights MENA: (www.insightsmena.com) Middle East-centric statistics site developed collaboratively with Google. Responsibilities: - Backend development using Java and Hibernate - Total UI frontend infrastructure and development using HTML5 and SVG plus jQuery - Google App Engine Cloud Environment Integration

Majored in Artificial Intelligence. Thesis project titled "Improving Multi-label Document Classification with Neural Networks". We dealt with the use of a neural language modeling technique called paragraph vectors to do document representation and use them to do text classification using LSTMs.