- Overall responsibility and accountability for implementing the Go-to-Market strategy and leading the technical engineering efforts involved in the design, implementation, and public demos of the company's cloud-based traffic management system - Trafluid.- Supervizing the end-to-end development lifecycle of state-of-the-art Multi-Agent Reinforcement Learning (MARL) algorithms operating at the core of our traffic management system. - Team management including hiring and onboarding new engineers, setting objectives for both the cloud system and RL development teams, and monitoring ongoing progress and performance.- Designing the Statements of Work and overseeing the timely delivery of deliverables to clients, including a US State Department of Transportation.

Working on Responsible AI (RAI) tools and Privacy Preserving Machine Learning (PPML) techniques (Homomorphic Encryption, Differential Privacy, Confidential Computing, Federated Learning and Secure Multi Party Computation among others) for better ML model transparency, fairness, security, privacy and accountability. This work led to the publication of two whitepapers:- A whitepaper titled "Putting Responsible AI into practice for your AI-based solutions" which provides a starter guide for data scientists with an overview of Miscrosoft's Responsible AI journey and the most prominent RAI tools, accompagnied with ready-to-use code for getting started with these tools. - A whitepaper titled "Implementing a Responsible AI Lifecycle for MLOps processes" which showcases an end-to-end walkthrough of a Responsible AI lifecycle for cloud-native applications based on MLOps, spanning a complete use case.Both whitepapers and the related code can be found in the Microsoft open source "Responsible AI Workshop" repository.

- Designing a cloud-based ML powered traffic management system using a microservices approach.- Developing the driver's mobile application for this cloud solution using Flutter.- Deploying a Kubernetes cluster for the system on GKE using Terraform.

Implemented a Deep Q-Learning algorithm for computing optimized traffic lights based on live traffic.

- Designed a simulator in Python that generates a network of vehicles on a map and runs the traffic control algorithm on it.- Used the simulator to find and resolve bugs in the core logic of the system.- Showcased the system to investors in Riyadh, Saudi Arabia in July 2019.

- Participated in the implementation of an infrastructure free traffic control scheme that leverages the presence of Vehicle-to-Vehicle (V2V) communication to produce intelligent virtual traffic lights.- Implemented in Java major pieces of the core logic running on the on-board units (OBU) mounted on cars.- Implemented in Python a coordinator used to launch all pieces of the system and manage their interaction.

-Helped students grasp the essence of differential and integral calculus during their exercise sessions -Coordinated with two calculus professors over two academic semesters and many student assistants to make these sessions suitable for 1st year students