Amel Mhamdi
Data scientist
38 years old
Driving License
Professional Status
Employed
Available
About Me
Experte intelligence artificielle diplômée d'un Doctorat en Intelligence Artificielle en 2016, je me suis spécialisée sur l'analyse des données ainsi que l'automatisation des processus de travail et la prédiction via les Framework de l'IA et les algorithmes de Machine Learning.
Éternelle optimiste et curieuse de nature, je m'intéresse à tout ce qui me passe sous la main.
J'aime apprendre de nouveaux métiers et outils liés au Data, IA et le cloud et je suis toujours prête à aider mon équipe pour atteindre nos objectifs !
Éternelle optimiste et curieuse de nature, je m'intéresse à tout ce qui me passe sous la main.
J'aime apprendre de nouveaux métiers et outils liés au Data, IA et le cloud et je suis toujours prête à aider mon équipe pour atteindre nos objectifs !
Experiences
Responsable projet Intelligence artificielle chez France Télévision Via Wewyse
France Television
Since July 2019
- Contribution
Cartographie et mise en place des projets IA pour le domaine de Media - Responsabilités :
• Participation à la préparation d’organigramme de l’unité d’intelligence artificielle transverse de France tv : DaIA
• Définition des cas d’usage pour la définition des projets IA rapide au déploiement.
Mise en place de la méthode agile aux seins de l’équipe.
• Benchmarking des solutions de marché pour la transcription (Speech To Text)
• Mise en place d’un corpus Ftv pour un service de Speech To Text adapté en différents accents françaises, ainsi qu’un outil de transcription (python : speech-recognition).
• Mise en place d’un service de traitement de langage naturelle (NLP) (python : Spacy).
• Encadrement d’une équipe de six étudiant de l’ECE pour la mise en place d’un outil automatique de dépouillement des scénarios (python : Django).
• Visualisation des données des débats d’élection municipale dans les différentes régions et à une échelle nationale.
• La coopération avec l’équipe Infrastructure pour définir les besoins en termes de services sur le Cloud AWS.
• Recueille directe des besoins d’utilisateur (rédaction, indexation, journaliste, Infographistes).
• Coopération avec l’EBU pour la mise en place d’outils media européen pour le Benchmarking des solutions Speech To Text et la mise en place d’une data Lake coopératives pour le test des différents algorithmes de machine Learning.
• Création d’appel d’offre et recrutement des profils data ingénieur
• Communication des résultats de recherche à des conférences internationaux. - Environnement Technique : Python, Postgree, Kibana, elasticsearch, Git, Docker, Jira, Confluence
Data Consultant
Edenred
Since September 2018
- Context:
As part of the digital transformation of the IS department of Edenred, I joined the team of Edenred Connect, responsible for the SOO and communication with the various APIs of the IS (CRM ...).
The application manages the authentication and transactions of the company's web services, employees and merchants, whether by card, mobile or on online platforms.
Team:
1 scrum master, 2 in, 5 developers
Project organisation:
Agile (Sprints of a week) - Contribution:
In charge of customer data, their use, their provision and their exploitation for the prediction of the fraudulent.
Set up a support automation tool: enable customers to collect problem solutions and get expert advice / knowledge about problems via chatBot, without having to call an agent support team. - Mission:
- Collection of need in close collaboration with the profession
- Management and administration of the database for Edenred-Connect
- Versioning Management under Azure DevOps
- Prepare data analysis in Azure machine learning and export from Azure machine learning
- Develop machine learning models
- Operationalize and manage Azure machine learning services
- Use other services for machine learning
- Documentation
- Code review
-Test and application recipe - Collection of need in close collaboration with the profession
- Technical environment:
- Agile SCRUM Methodology
- C #
- Git, JIRA
- AzureDevOps
-Azure machine learning studio - Agile SCRUM Methodology
Expert ChatBot
BNP Personal Finance
June 2018
to August 2018
- Context:
BNP Paribas Personal Finance, a subsidiary of the BNB Paribas group, offers a full range of personal loans available in stores, in car dealerships or directly from customers via its customer relations centers and on the internet.
As part of the credit simulation, I joined an Agile team: Digital Lab, in charge of creating a credit simulation chatbot.
Team:
1 scrum master, 4 in, 3 developers
Project organisation:
Agile (Sprint of a week) - Contribution:
- Establishment and development of the Bot architecture.
- Collection of need in close collaboration with the trade (PO)
- Implementation of unit tests and integration test
- Chatbot functional tests
- Establishment and development of the Bot architecture.
- Technical environment:
- Agile SCRUM Methodology
- Crafts: Unit Testing, TDD
- language: Nodejs
- Agile SCRUM Methodology
Expert Artificial Intelligence / Machine Learning
Natixis
October 2016
to May 2018
- Context:
As part of the digital transformation of BPCE, I joined an Agile team, responsible for creating an application that predicts the re-socialization of Front Office reconciliations. The team includes collaborators who have never practiced Agile methods,
So I upgraded the team members to Agile methods to improve their autonomy and efficiency.
Development of a tool based on Machine Learning and artificial intelligence frameworks to automate the creation of reconciliations.
Maintain an active watch on the theoretical and technological topics of data science.
Create a ChatBot for support automation of a Front Office reconciliation application.
Team:
1 Scrum Master, 7 Developers, 1 PO Proxy (which links the clients to the team), 3 MOA
Project Organization:
Agile (2 weeks sprints).
Part of the team is in Portugal (3 developers), the rest of the team in Paris. - Contribution:
Set up a prediction tool for creating bank reconciliations via the strategies of artificial intelligence (CSP) and machine learning (classification algorithms).
Functional needs analysis in close collaboration with the business.
Implementation of the ChatBot architecture.
Preliminary study of the different existing frameworks for the creation of a chatbot.
Implementation of different technologies to use.
Chatbot functional tests - Mission:
Knowledge in market finance (options market, stock market and bonds)
Design, analyze and maintain databases,
Create, calibrate quantitative prediction models using Machine Learning and Artificial Intelligence (CSP) techniques
Synthesizes and communicates the results of mass data analyzes to extract a decision-making knowledge
Enhance existing algorithms and develop new algorithms for creating automatic reconciliation of Front Office deals.
Creation of ChatBot architecture.
ChatBot integration with a Front Office reconciliation web application.
data trapping and visualization - Results:
Development of a ChatBot for functional and technical support automation
Creation of a predictive tool for creating bank reconciliations - Technical environment:
Computer programming: Python (scikit-learn, Numpy, SciPy, Pandas, JSON, CSV, XML, CSP, mysql.connector, pytest, unittest)
Algorithms and database management: SQL
Agile SCRUM Methodology
Crafts: Unit Testing, TDD
link of the Certification: https://www.youracclaim.com/badges/332ca581-f00f-4f16-8b91-7c32de918c3a/public_url
Context:
Although the problem of graph editing (CE) has been defined since 1964, it is still very much studied. One of the reasons is that the graph editing problem occurs in many applications such as, for example, species analysis, a problem in the field of genetics. The difficulty is to choose a representation that guarantees an effective treatment of this problem. Several representations have emerged, such as the one that uses: the logic of the propositions (SAT) and its extensions (2-SAT, 3-SAT, MAX-SAT, ...), the formalism of the problem of satisfaction of the CSP constraints.
Contribution:
Our contribution in the field of the classification under constraints is varied according to two main axes:
Results:
the goal is to propose a new approach to deal with graph editing. This approach is based on the formalism of the problems of satisfaction of valued constraints.
This formalism provides a rigorous framework, more general than that of the CSP and especially powerful algorithms to efficiently solve combinatorial problems.
Publications:
Technical environment:
ToolBar2, C ++, Python
Although the problem of graph editing (CE) has been defined since 1964, it is still very much studied. One of the reasons is that the graph editing problem occurs in many applications such as, for example, species analysis, a problem in the field of genetics. The difficulty is to choose a representation that guarantees an effective treatment of this problem. Several representations have emerged, such as the one that uses: the logic of the propositions (SAT) and its extensions (2-SAT, 3-SAT, MAX-SAT, ...), the formalism of the problem of satisfaction of the CSP constraints.
Contribution:
Our contribution in the field of the classification under constraints is varied according to two main axes:
- Methodological: we propose methodological approaches adapted to the problems of graph editing and the problem of classification by deletion of edges.
- Technique: we introduce new techniques (heuristic search algorithms, global constraints) to fill some gaps identified thanks to our methodology.
Software: The implementation of these techniques in a constraint solver facilitates their diffusion and their implementation in various applications.
Results:
the goal is to propose a new approach to deal with graph editing. This approach is based on the formalism of the problems of satisfaction of valued constraints.
This formalism provides a rigorous framework, more general than that of the CSP and especially powerful algorithms to efficiently solve combinatorial problems.
Publications:
- "Constraint programming approach to the Zahn's decision problem". In the Proceeding of the International Symposium on Distributed Computing and Artificial Intelligence (DCAI 2014), Salamanca, Spain, June 2014.
- "A constraint programming approach to the cluster deletion problem". Proceeding of the 13th Scandinavian Conference on
Artificial Intelligence Halmstad, Sweden, November 5-6, 2015
- "Classification of genes based on weighted PUC".
The Eleventh French Journées de Programmation by Constraints, (JFPC 2016).
- "LDS vs FPT method for cluster deletion International conference of artificial intelligence tools.
28th IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2016, San Jose, CA, USA, November 6-8, 2016.
- "A constraint programming approach to the cluster editing problem". International Journal of Imaging and Robotics (ISSN 2231-525X), August 2015.
Technical environment:
ToolBar2, C ++, Python
Topic: transforming component-oriented meta-models (ACME) to a service-oriented meta-model
Publications:
Publications:
- Towards an IDM Approach of Transforming Web Services into ACME Providing Quality of Service.
Seventh International Conference on Semantics Knowledge and Grid (SKG 2011), Beijing, China, October 24-26, 2011.
- Towards a transformation of composite web service with QoS extension into ACME\Armani.
iiWAS'2011 - The 13th International Conference on Information Integration and Web-based Applications and Services, 5-7 December 2011, Ho Chi Minh City, Vietnam.
Programming engineering (internship: creation of a payroll software based on the digital borrowing with FoxPro)
Machine learning, data mining, text mining, graph mining, data analysis, data visualization, constraint programming
Decision Tree Preparation, Natural Language Processing, Robot / Human Conflict Resolution
Intellectual curiosity
statistical analysis and calculation of data
Unstructured data management
data processing and visualization
Visual Studio 2012-2103-2017: c#( asp .Net et .Net core 2.x)
SQL SERVER 2012-2014,2017, Sybase, Oracle
TFS ,GIT, XLDeploy, Jenkins, Jtrac
jupyter : python et R
Interests
Sports
- fitness
- Yoga
Going out
- schedule outings with workshops for children
Travel
- Discover the habits and history of the different countries
