MAJID ALAMDARI

Analyzed and reviewed documentation of IVECO diesel engine, so as to understand how each component works.

Received the data streaming from Databricks in the mf4 format, using data mining techniques as well as
automated tools to clean and process the data, making it ready to be applied to machine learning models.

Selected the optimal set of features for each component, enabling the construction of an optimized deep learning predictive model.

Defined an index for each component called the health index, to prevent, diagnose, and detect failures of the
components.

Achieved an almost perfect loss on the test set, demonstrating an MAE of 1% within the variable range.

Experienced in monitoring cloud infrastructure across distributed systems using Azure, Datadog, Lens, Linux,
Akamai, and the Bash shell.

Experienced troubleshooter adept at solving complex technical issues, managing tickets from international clients under tight deadlines, addressing issues related to bugs and shortcomings across various projects, utilizing Jira and Confluence.

Worked in a company providing various services, from Internet to satellite channel subscription to customers.
Managed customers on each terminal to reduce traffic, writing automated Python scripts using the Python REST API library asd FastApi framework to streamline the process automatically.

SKILLDO Project :

Conducted a research project in collaboration with sports and parallel sports medicine, focusing on
comprehensive analysis of sporting activities from individual player performance to team dynamics.

Specialized in match analysis, employing Python software and leveraging specific libraries such as 'METRICA' tailored for football event analysis.

VEDO Project :

Collaborated in a project focused on the industrial production of EGR solenoid valves with an integrated diagnostic system.

Collaborated with the machine learning team to conduct statistical analysis of input data, defining reliability and alarm ranges for the diagnostic system.

Developed an algorithm for forecasting, enabling the calculation of valve efficiency even when direct data
connection was unavailable.

Initially developed algorithms in R software and later transitioned to Python, utilizing the KERAS library for model development.

Achieved an almost perfect loss on the test set, demonstrating an MAE of 1% within the variable range.