Sara LAOUAR

Experience

• 

  HOTEL BONNE NOUVELLE

  May 2018 - Jul 2018

  Receptionist

  * Served breakfast* Welcomed, customers, answered telephone, check-in/check-out * Responded to the needs of customers

• 

  Private Medical Imaging Center

  Jun 2019 - Jul 2019

  Medical Secretary

  * Welcomed and helped patients, made appointments, entered patients personal information* Answered telephone and made appointments* Helped the doctor by drafting medical reports

• 

  Freelance

  Dec 2020 - Dec 2020

  Mathematics and chemistry tutorials
• 

  Pharmacologic and Toxicological Chemistry and Biochemistry Laboratory, Nano Bio-Spectroscopy team

  Jun 2021 - Jul 2021

  Biomedical sciences Bachelor degree Internship

  * Synthesis of nanoparticles (in situ = in solution) and synthesis of hybrid microgels containing gold nanoparticles, * Incorporation/trapping of probe (or drug) in the gels and incorporation of gold nanoparticles into gels* Study of probe (or drug) delivery/release by hybrid materials: passive delivery (without stimuli) and active delivery (pH or temperature change).

• 

  Private Medical Imaging Center

  Jun 2022 - Jul 2022

  Immersion Internship

  * Known software used : Hologic Horizon WI for osteodensitometry, Evolucare imaging RIS-PACS for organize, share and archive medical data and technologies DXA and HIRIS RF43 (high resolution sensors, low noise and fast).* Known to use Hologic Horizon WI software (calculate bone density and risk fracture)

• 

  BeHIT

  Jun 2023 - Sept 2023

  Digital health Internship

  * Worked on the development of an interoperable data architecture for clinical data.* Worked on a solution capable to collect data, adjust the format on OpenEHR (.json), clean the data and train a data model and test a model using Gradient Boosting and GridSearchCV methods.* The works include : - data collection, on different formats .CSV/MySql/.XML - data analysis, cleaning the data, analyzing outliers - data science, on Python developing data models based on Gradient Boost and recommandator algorithms* Improved the model : surpass the sensitivity with 88% instead of 65%* Taken part in a scientific article Show less

• 

  Worldline

  Mar 2024 - Sept 2024

  Product AI Prompt Engineer

  Onboarding:- Training on : RAG, fine-tuning, and generative AI.- Training on SCRUM methodology.- Finance Training: Introduction to card payments, end-to-end SEPA training, commercial acquiring, and an introduction to e-commerce.Projects:AI Solutions Development: Implementing AI solutions (Generative AI, Fine-tuning, RAG) to optimize document management and internal search. Analyzed limitations of current systems (Confluence, JIRA) and handled LLM token constraints. Created an AI Assistant for internal documentation, enhancing search accuracy and initiating AI integration at Worldline. RAG:- Designed a vector search system with document indexing and similarity scoring.- Utilized Langchain for LLM question/answer systems.- Identified and discussed issues and solutions. Fine-tuning:- Researched manual and automated methods using Vertex AI.- Handled data acquisition and formatting (JSONL).- Evaluated robustness with varied data volumes.- Fine-tuned models (Bison, Gemini) with public and confidential data. AI-Powered Internal Documentation Assistant:- Led a project to improve internal documentation management using artificial intelligence.- Linked PDFs with LLM for TAP questions on Confluence.- Improved prompts and updated restriction files.- Extended the scope beyond the initial Confluence page.- Resolved issues and provided solutions. Hackathon for Project Management and API Assistant Development Using LLM:- Led in creating an AI-based integration assistant for Worldline APIs.- Simplified API integration, provided call examples, simulated error cases, and facilitated test environments.Technical Skills:- Project Management Tools: JIRA- Development Methodologies: Scrum- Version Control: GitLab- Development Environment: VS Code- Programming Languages: Python Show less

• 

  Square Management

  Jan 2025 - now

  Data/AI Consultant

  Design and Development of an AI Training Program for Consultants and Practitioners, structured into three levels:- Level 1: Introduction to AI and Gemini Advanced - Level 2: Mastery of Prompting and Language Models - Level 3: Optimization and Integration of LLMs Participation in the Preparation of Commercial Proposals for AI-Related Projects (BPCE, CCF)R&D Project - Benchmark of RAG Solutions:- Objective: Conduct a comprehensive benchmark of algorithms and solutions to develop a pipeline based on the RAG architecture.- Focus: Identify the most efficient and suitable solution for each component of RAG (OCR, embedding, retrieval, ranking, reranking, generation, moderation) to meet client needs.Impact: Assist consultants and clients in selecting the best technologies for their RAG projects.- Context: As part of this project, the evaluation of the best solutions to optimize client pipelines was undertaken. This began with an analysis phase of OCR solutions (PyPDF, Unstructured, LlamaParse), including user testing through a Streamlit interface and business documents. My work specifically focused on integrating Azure's Document Intelligence to complete this first phase before extending the evaluation to embeddings.- Contributions:*Implementation of Azure's Document Intelligence: Developed a Python script to convert PDFs into plain text (markdown, tables, images) using Azure's Document Intelligence API. Managed large volumes of data and optimized performance.*Participation in Organizing and Analyzing User Tests: Organized tests with consultants on around a hundred relevant questions per document, using specific evaluation criteria. Conducted both quantitative and qualitative analysis of results to identify the strengths and weaknesses of each OCR solution (PyPDF, Unstructured, LlamaParse, and Azure Document Intelligence)*Research and State of the Art on RAG Components: Explored chunking methods, embedding models, and retrieval techniques to optimize RAG Show less