Abdelbari Bouzarkouna

I lead a team of data and softwre sienctist, our tasks to bring AI and ML into manufacturing 4.0 to generate more revenue streams and reduce costs.

During my current experiences we work on several use cases so I had the opportunity to manage and contribute all stages of a data science project:

• Scopping the business needs and helping the project sponsorship to estimate the Capital expenditures and the return on investment.

• Prepare and statistically analyze the data in order to properly select and create features better understand the problem through data.

• Select the right algorithm of the Modellintion of the problem with the appropriate AI methods (ML, DL or CV models).

• Validate the solutions with the business based on statistical metrics.

• Improve the models and go for an MVP (Minimum Viable Product)

• Structure the code as a pipeline flow (kedro framework) and version it with git

• Made the cloud architecture of the projects and data pipelines to mutualize use cases and ease deployment on the cloud envirmenet (Azure Cloud)

• Creation of CI and CD peplines to guarantee an agile and mlops mode that aims to deploy and maintain machine learning models in production reliably and efficiently.

• Deploy the solutions for real time recommendations on an event based architecture into different types of environment: development, pre-production and production environment using containerized architectures including Docker and Kubernites.

• Finally send those live recommendations back to the factory so that they can be displayed and acted on by an operator on the field.

• Monitoring the data science chain: infrastructure, model performance, data versioning.

• Documentation and user support in order to get feedback and future improvements.

Technologies : Python / Docker/ Kubernetes / SQL / Queues/ Azure Blob Storage / PowerBI / Azure Devops (CI/CD)/ Azure Synapse Analytics/Data Lake Analytics/HDInsight .

Methodology : Agile Scrum / MLops

As a member of the AI team I worked on several projects:

1- A Computer Vision & Deep learning projects (Real-time facial recognition, object detection).

2- A NLP projects (Classification, Sentimental analysis).

Used tools :

• Python : Pandas , Numpy, Sci kit, Nltk, Mpld3, TensorFlow, OpenCV, Dlib, Pytorch.

• Azure Cloud: AKS / Azure Synapse Analytics/ Azure devops.

Developing analytical methods for unconventional and conventionnal assets: Forecasting on time series using machine learning/ Deep Learning for business purposes. As a part of a team I helped in:

• Developing of a Power Bi interface for the exploration of raw data, detection of outliers, management of missing values.

• Adapting and optimizing the algorithm (LSTM , Random forests ...)

• Studying different deployment and production solutions.

• Developing web application to interact with the model and visualise the results.

• Deploying the model into production. Tools:

• Python : Pandas , Numpy, Sci kit, Keras, Tensorflow, Tensorboard, Flask, Apache Airflow.

• Azure Cloud: Azure Synapse Analytics/ Data Lake Analytics/ Azure Stream Analytics.

• Web: Javascript / Html/ Bootstrap 4.0/ D3.js/ Plotly.js .

• BI: Power

• Project management: Jira (agil Kanban) / BitBucket (git) .