Who am I?

I am Mohamed Chelali, researcher at the crossroads of computer vision and artificial intelligence. Passionate about technological innovation and solving complex problems.

Keywords : Video Processing, Tracking, Violence Detection, Satellite Image Processing, Image Time Series, Computer Vision, Remote Sensing.

Resume in English, CV in French

Timeline

[Aug 2024 - Feb 2025] - Data Scientest

Machine Learning Ops

Training in technical skills for automation, deployment, and continuous monitoring of machine learning models in production

[2022-May] - Jellysmack

AI Applied Scientist

Designed, prototyped and put into production an engine with artificial intelligence to summarize a video.
Internal consultant for the description of images in sentences, tags and other descriptors.

[2018-2021] - Université Paris Cité

PhD in Computer Science

Thesis title : Image time series analysis involving spatial and temporal information

New constellations of satellites acquire images with a high spatial, spectral and temporal resolution around the world leading to Satellite Image Time Series (SITS). During my thesis, our objective is to analyse the image time serie in order to extract features that enrich the orginal data and perform an application, such as a classification, by maximazing the accuracy
Advisors: Nicole Vincent et Camille Kurtz

PhD defense page

[2016-2018] - Paris Descartes University

Master's degree in Computer Sciences - Image et Plurimedia field

The master's was validated with a research internship at the LIPADE.
Advisors: Nicole Vincent, Camille Kurtz.

[2015-2016] - El Bachir El Ibrahimi University

Master 1 in Computer Science - Networking and Multimedia field

[2012-2015] - El Bachir El Ibrahimi University

Bachelor's degree in Computer Science - Image Processing field


Research

, .
[PDF] [DOI] [URL]

Journals

Books

Conference and Workshop Papers

National conference and Workshop Papers

Presentation

Thesis


Project

Temporal stability
...

Spatio-temporal hand-crafted features extraction that measure the stabilty. The proposed method is based on a compression algorithm named Run Length Encoding.


Temporal stability code available on github
Phd Thesis Template
...

This LaTeX template is designed to streamline the creation of a professional and well-structured PhD thesis. It adheres to academic formatting standards while remaining highly customizable to meet the specific requirements of institutions. This template provides predefined environments for chapters, sections, figures, tables, and citations, with additional styling for the title page, table of contents, and appendices.


Template is available on github
Template is available on overleaf

Teaching

I had a teaching assignments during of my Phd research at the UFR Mathématiques et Informatique of Université Paris Cité

I had a teaching assignments at the IUT of Université Paris Cité

Tutorials