Skip to Content

AI4FINOPS

Owner : Mouna Ben Mabrouk
April 26, 2024
  • Abstract
  • Technology

AI4FinOps is an automated platform aiming at harmonizing collaboration between DevOps and Finance teams to optimize expenses related to cloud infrastructure. AI4FinOps enables a comprehensive understanding of costs by linking expense data to the business. It establishes compliance strategies and tags and calculates custom rates for better financial management. The platform provides timely expense data to enable real-time decision-making. It identifies anomalies almost instantly and effectively communicates with relevant stakeholders for increased responsiveness. AI FinOps optimizes cloud resource usage by resizing underutilized resources. It covers not only compute but also databases, storage, and managed services, ensuring elasticity and scalability. The platform balances the use of different pricing models, such as on-demand, reserved capacities, spot market, etc. This maximizes savings while considering fees and flexibility. AI4FinOps offers personalized plans based on each company’s specific needs. This balances costs and engineering efforts with overall savings, while setting a minimum savings threshold for effective resource management.

ChatBot, NLP, LLM, XAI, DevOps, Deep Leaning and Machine Learning

ELIT

Owner : Ines Ben Kraiem
April 26, 2024
  • Abstract
  • Technology

Project management relies on a rigorous methodology, integrating defined practices, logics, and processes, aimed at orchestrating optimal planning, development, and control of a project throughout its lifecycle to ensure its success. However, given the numerous variables influencing projects, especially in the information technology ecosystems of businesses, determining the ideal methodology for a specific project becomes complex.
The ELIT project aims to address this complexity by developing and implementing a decision-making approach. This approach, tailored to the unique characteristics of each IT project, will identify the most relevant project management methodology. Additionally, the ELIT project aims to evaluate and predict the viability of projects from their inception through the use of artificial intelligence. This early anticipation of potential obstacles and challenges will provide the opportunity to take corrective measures or abandon projects with high risk.
To address these challenges, we utilize machine learning algorithms and generic artificial intelligence. These tools facilitate and automate project verification and validation from its initiation, through feasibility studies and risk analysis, to the selection of the most suitable project management method.

Project management, Agile, decision-making, Machine learning, Generative AI

TDA

Owner : Bishal Poudel
April 26, 2024
  • Abstract
  • Technology

The TDA project is located halfway between research in applied mathematics and information theory. The objective of this project is to develop TDA methods while connecting them to other fields such as data science, quantum physics (e.g. topological matter, exotic phases, experiments like ARPES, etc.), computational chemistry, biology, and medical imaging. Firstly, we would like to implement efficient TDA methods such as persistence homology and mapper in order to analyze data from the French electricity network. Thereafter, we would like to extend this project to other use cases, along with the development of new techniques.

TDA, Information theory, Algebraic Topology, Data Science

GREEN IT

Owner : Mouna Ben Mabrouk | Co-Owner : Md Siddiqur Rahman
April 26, 2024
  • Abstract
  • Technology

Today, digital activity is contributing to an increase in overall energy consumption and CO2 emissions. To cope with this issue, different research works have been carried out to optimize digital energy consumption. However, existing solutions for optimizing digital energy consumption are proving ineffective due to the complexity and heterogeneity of IT infrastructures. The aim of this project is to develop innovative solutions for effectively monitoring and optimizing the energy consumption of an IT infrastructure, whatever its architecture (Cloud or On Premise) and its complexity (diversity of equipment and tools).

Green IT, Cloud computing, machine learning

MEDIBOT

Owner : Sogeti Labs
April 26, 2024
  • Abstract
  • Technology

According to several studies, social isolation increases the risk of cardiovascular disease, Alzheimer’s and other conditions. This applies not only to the elderly, but also to children with neurological disabilities. The solution to this problem often lies in admitting this population to specialized care facilities (nursing homes, wards for the neurologically handicapped) for long-term stays. However, the lack of nursing staff and the remoteness of these patients from their loved ones are not conducive to quality care. To address this issue, we plan to propose an autonomous communicating robot that will improve the well-being of these patients by 1) establishing a means of communication between patients and their families, to help maintain strong ties; 2) a care support function to assist caregivers in the task of periodically monitoring the patient’s condition, to enable rapid management in the event of a deterioration in health status.

Computer vision, IA, robotic, medical diagnostic

MYSPADE

Owner : Amine Ferdjaoui | Co-Owner : Emmanuelle Kelodjoue
April 26, 2024
  • Abstract
  • Technology

MySpade addresses an internal challenge within the group, focusing on enhancing human resource management, particularly addressing turnover and attrition. It aims to develop a tool to foster the long-term integration and inclusion of new employees within Capgemini, while also mapping their career trajectories based on existing profiles within the company.
The theoretical objectives include proposing a decision support system applicable to the corporate world, developing a new information extraction system, and exploring various recommendation approaches using fast and effective NLP techniques.
Economically, the project aims to support the recruitment process and tailor training proposals to employees based on their profiles, and implement an internal mentoring system to assist employees in mapping their career paths and reducing turnover.

NLP, Causality graphs, clustering, recommendation systems

COGNICITY

Owner : Abdul Gizzini
April 26, 2024
  • Abstract
  • Technology

Smart cities are urban environments that leverage advanced technologies and data-driven solutions to improve the quality of life of their inhabitants. Applications of these cities, such as object detection for autonomous vehicles and license plate recognition, aim to optimize the use of available resources to ensure both efficiency and safety. Despite the diversity of solutions available in the literature to tackle these applications, they often face problems of limited performance and lack of interpretability. This situation compromises their suitability for real-world, secure deployment. To address these challenges, we present Cognicity, a framework based on artificial intelligence. Cognicity draws on advanced data analysis technologies to design AI models that are both accurate and easily interpretable. By providing solutions that are both high-performance and transparent, Cognicity aims to improve the relevance and security of real-world deployments in the context of smart cities.

Smart City, Computer vision, Object detection, Artificial intelligence (AI), Explainable AI (XAI)

TOMORROWWAN

Owner : Abdul Gizzini
April 26, 2024
  • Abstract
  • Technology

5G networks will have to support new multi-industry applications and services with radically different requirements. In order to deploy these various services within a network, the TWAN project will leverage SDN and NFV technologies to create and manage isolated network partitions (slices) on a shared WAN (Wide Area Network) infrastructure. The project will propose a software platform for orchestrating a WAN network. We are looking for a set of solutions to automate this network and optimization algorithms for the dynamic reconfiguration of multi-domain and multi-technology WAN partitions. The main idea will be to optimize routing and service deployment using AI technology. The solution will also have to take into account the following aspects: minimizing energy costs and latency, and optimizing resource use. We’ll also be working on an intelligent top layer for decision support. The resulting network will be a Knowledge defined network (KDN) where the logic of AI-based decisions will be explained, ensuring robustness and confidence in the solutions developed.

Network Automation, Traffic Engineering, SDN, SD-WAN, AI, XAI

GMS

Owner : Adel Taleb
April 26, 2024
  • Abstract
  • Technology

The GMs project aims to develop models for utilizing standard quality video streams using classification methods and anomaly detection for the prediagnosis of neurodegenerative diseases in newborns.
The objectives are as follows:

  • Pose estimation on a newborn’s motor skills
  • Study of a newborn’s normal spontaneous motor skills according to age
  • Detection of pathologies using non-invasive and inexpensive tools.

AI for Healthcare, Video processing, Clinical trials

TNT

Owner : Wissam Mammar Kouadri
April 26, 2024
  • Abstract
  • Technology

TNT is a project that is interested in improving commercial and HR tasks in a company by introducing artificial intelligence. In TNT, we work on two main research axes. The first research axis concerns conversational recommender systems, where we aim to facilitate the process of matching consultants and missions by providing an interface that uses natural language prompting. The second axis concerns human-machine interaction, where we study the effect of conversational interfaces on user experience

NLP, HCI, data quality