Mohamed T.

➔ Implementation of optimization algorithms (PSO) for traffic light management.

➔ Participation in R&D on new approaches for traffic management based on reinforcement learning PPO.

➔ Setup and monitoring of self-hosted Kubernetes deployments from scratch.

➔ Implementation and automation of backup strategy for PostgreSQL databases.

➔ Rewriting code for various IoT services, Django for embedded and user applications.

Environment: Python, Kubernetes, Reinforcement Learning, RLlib, Data-Science, MicroK8s, MLOps, MongoDB, Postgresql (Geo-Data)

Data Science :

➔ Develop embeddings for the knowledge graph project, which were used to improve the segmentation of different user profiles using clustering models on processed time series data.

➔ Contribute to the development of a more advanced insurance scoring model for drivers, to better define their pricing based on their driving habits.

Machine Learning :

➔ Optimize and develop custom models for internal and external use.

➔ Implement a strategy to detect when models drift and retrain them with new data.

➔ PoC of an end-to-end automation strategy (from code commit phase to model production deployment).

Environment: Windows, Databricks, Neo4j, Tensorflow, AWS, Snowflake, Sklearn, github, Python, Scala, MLOps

➔ Research best techniques for managing the cloud and network needs of robotics applications (requirements) that are agnostic to the cloud provider and network.

➔ Define the architecture of the proof of concept (PoC) for cloud and network using a classification model.

➔ Build the PoCs and make them usable using Kubernetes, Python, C++ and Nokia's 5G simulator.

➔ Write technical documentation for the PoCs as well as the methodologies implemented.

Environment: Linux, C++, Python, Sklearn, Kubernetes, Flask, Real-time, Docker, Kubernetes, Helm