Jonathan DEKHTIAR

PhD. in Deep Learning Solutions for Manufacturing Applications.

University of Technology of Compiegne

Thesis directors : Alexandre DURUPT (UTC) and Dimitris KIRITSIS (EPFL).
Co-Supervisor : Matthieu Bricogne (UTC)

PhD Subject : Deep Learning Applications for Vision Based Systems in the manufacturing industry.

During my PhD. I had the opportunity to work on numerous fascinating projects and challenges:

• DMU-Net: an open-source dataset oriented toward the manufacturing industry to perform recognition task in industrial environments.
• DICE - DMU Imagery Classification Engine: a deep convolutional neural network to perform 2D and 3D data classification concurrently.
• Manufacturing Visual Security Kit: a Deep Learning toolkit to insure that no object can be damaged oo any human can be hurt during a manufacturing process.

Guest Researcher - Deep Learning Scientist

Technical University of Berlin and Fraunhofer IPK

Engineering Master in Computer Science and Data Science

University of Technology of Compiegne

TOEIC : 975/990

Test of English for International Communication

Developped and organized by the US-based company : ETS, the TOEIC Listening and Reading Test is a two-hour multiple-choice test consisting of 200 questions evenly divided into listening comprehension and reading comprehension. Each candidate receives independent scores for listening and reading comprehension on a scale from 5 to 495 points. The total score adds up to a scale from 10 to 990 points.

A Score of 975 corresponds to a C1 Level in the European Scale.

Two Erasmus: Hamburg and Vienna

Technische Universität Hamburg Harburg

I have been studying for one semester at each of these university:

• Technische Universität Hamburg Harburg
• Technische Universität Hamburg Harburg

Data Scientist and C++/Python Developer

DeltaCAD

I have done a 6 months Internship inside the company DeltaCAD under the supervision of Harvey ROWSON.

I have designed and developed a software application to perform mechanical parts recognition and segmentation. The core is developed in C++ which is exposed using a Django as a REST API hosted on AWS (Amazon Cloud).
During this project I benchmarked major machinelearning algorithms and boosting techniques to improve the obtained results.

Founder, Head of the organisation and Licence Owner

TEDxUTCompiegne

We are already today the core of innovation and creativity and tomorrow will be too late. It's high time to act in order to shape the world and not just to cope with it. My engineering background relies not only on scientific topics but also on open-mindedness and assumption of responsibility. With the aim of bringing to the future engineers the keys to the great social and technological challenges, we have built TEDxUTCompiegne, and the first event took place on the Tuesday January, 20th 2015.

Photos : Album Flickr
Videos : Playlist Youtube - TEDxUTCompiègne - January 20, 2015
Official Website : TEDxUTCompiegne.com


Our thema : “Observing, Imaginating, Achieving”.

Three actions verbs that together describe the innovation process. Three action verbs that reflect a willingness to act and understand. Three verbs that lead my way. No other thema could have transmit better my determination to engage myself in the social and technological challenges of tomorrow.

Speakers :

Business Automation Analyst

Valeo

I have done a 6 months Internship inside the Information System Shared Service of Valeo Confort and Driving Assistance under the supervision of Marc Laurent.

In order to prepare a SAP migration, I have developed a tool to clean the data from the SD Module. The tool was a developed in VBA in Excel and performed a Statistical Analysis on clients’ data from our own system from to those in the global SAP system. The goal was to match the clients of the two systems and identify those who didn’t exist in the global system and needed to be created.