R&D Project:

• The aim is to create a chatbot tailored to the needs of the business by using their documentation and integrating it into an LLM in order to specialise it.

Creation of a Retrieval model

• Data collection and preparation: Structuring business
• Documentation for accessibility and interpretability.
• Choosing the model architecture: Using LongChain for query processing and Chroma for indexing and retrieving relevant documents.
• Model training: Adapting the model to the specific characteristics of the business data to optimise the correspondence between queries and documents.
• Indexing: Transforming and indexing documents with Chroma for fast, efficient searching.

Integration of the Retrieval model with the LLM

• Model coupling: Link the retrieval system to Mistral to enrich LLM responses with specific business data.
• Input pre-processing: Simplify or reformulate user queries for better interpretation by the retrieval system.
• Post-processing of responses: Ensuring that responses comply with business requirements, possibly with reformulation for clarity.
• Integration with existing system: Make the chatbot accessible via the usual user interfaces, enabling fluid interaction.

R&D Project:

• As part of an R&D project dedicated to aviation safety, my role in collaboration with Airbus was to design and implement an innovative AI solution. The aim of the project was to use natural language processing (NLP), classification and time series analysis to automatically generate complex and varied aviation incident scenarios, thereby improving incident prevention and response strategies.

Creation and optimisation of Machine Learning models:

• Development of custom models using PyTorch, TensorFlow and Scikit-Learn for increased accuracy in scenario generation.
• Implementation of advanced algorithms for time series processing, improving the prediction of aeronautical events.

Advanced NLP research:

• Conducting exploratory studies to identify new NLP approaches applicable to aviation safety. Collaboration with experts in the field to validate the approaches and techniques used.

Anomaly detection on time series:

• Study and creation of a variational auto encoder model for anomaly detection between time series from normal and abnormal collar parameters (using Sagemaker on AWS to run high-performance models).

Development of a Functional Demonstrator:

• Design of an interactive user interface (HTML/CSS/JS) integrated with a Flask backend to demonstrate the system's capability.
• Development of a dashboard for real-time visualisation of data analysis results.

Data exploration and cleansing:

• Advanced use of Tableau for visual insights and Python libraries (pandas, NumPy, Matplotlib, Seaborn) for data processing and analysis.
• Development of robust data pipelines to ensure the integrity and quality of incoming data.

Performance Analysis and Model Optimisation:

• Implementation of performance metrics to evaluate and refine models.
• Continuous adaptation of models based on feedback and test results.

Process Automation:

• Creation of scripts to automate repetitive tasks, increasing operational efficiency.
• Integration of MLOps solutions for easy deployment and maintenance of models.

Missions:

• Design and development of electric flight control systems for aircraft.
• Selection, characterization, and integration of composite materials (wetlayup, prepreg) for aircraft's internal and external structures.
• Composite parts design, considering aerodynamic, structural, and safety constraints.
• Calculation and analysis of structures to ensure their strength and reliability.
• Construction and optimization of specialized ovens for composite materials.

Skills:

• Aircraft Systems and Avionics
• Composite Materials Engineering
• Structural Analysis and Design

My missions:

• Utilization of industrial data for making informed strategic decisions.
• Optimization of production processes by identifying existing weaknesses.
• Use of visualization tools to transform data into actionable insights.
• Development and implementation of solutions to increase efficiency and profitability.
• Continuous improvement of industrial performance by reducing costs and fostering process evolution.
• Establishment and execution of training programs to enhance staff skills.

Skills:

• Process analysis and problem-solving.
• Data mining and visualization.
• Continuous improvement and performance management.

Mission:

• Development of a CATIA tool.
• Creation of MECAMASTER for the design office, a tool enabling large-scale dimensional chain management.

Skills:

• CATIA Development
• Software Development
• Dimensional Chain Management

Curriculum:

• Python Programming: Mastery of Python programming, data structures, and libraries
• Data Manipulation and Analysis: Advanced techniques for data cleaning, preprocessing, and exploratory data analysis
• Machine Learning: Proficiency in machine learning algorithms and techniques
• Deep Learning: In-depth knowledge and practical experience with neural networks, deep learning architectures
• Statistical Analysis: In-depth understanding and application of statistical concepts
• Big Data and SQL: Proficient in working with large datasets and mastering SQL for querying databases
• Data Engineering: Advanced knowledge of data pipelines, data storage, and managing data flows
• Real-world Projects: Extensive hands-on experience in working on complex data science projects from start to finish

Curriculum:

• Engineering Fundamentals: Learning the basics of engineering, fundamental sciences, and principles of industrial engineering
• Computer Science and Programming: Mastering computer science and programming, including commonly used languages in the field, data structures, and algorithms
• Industrial Project Management: Techniques for managing industrial projects, planning, monitoring, and coordinating resources
• Analysis and Modeling Methods: Learning analysis and modeling methods to solve complex problems in industrial engineering
• Automation and Systems Control: Studying automated systems and control techniques to optimize industrial processes
• Information Systems and Databases: In-depth knowledge of information systems, database management, and information technology for process optimization
• Industrial Engineering and Logistics: Understanding principles of industrial engineering and logistics to enhance supply chain and production efficiency

Curriculum:

• Mechanical Design: Learning principles of mechanical design, drafting, and computer-aided design (CAD) tools.
• Materials and Manufacturing Processes: Understanding different materials used in engineering and various manufacturing techniques
• Mechanical Systems and Components: Studying mechanical systems, components, and mechanisms used in various industries
• Thermodynamics and Heat Transfer: Learning the basics of thermodynamics, heat transfer, and energy systems
• Fluid Mechanics: Understanding fluid behavior and principles of fluid mechanics in engineering applications
• Control and Automation: Introduction to control systems and automation in manufacturing and industrial processes
• Quality Management: Exploring quality control and management principles to ensure product quality and process efficiency
• Industrial Project Management: Techniques for managing engineering projects, planning, and coordinating resources
• Computer-Aided Manufacturing (CAM): Utilizing computer software to aid in manufacturing and production processes