About Agnès
- Object detection, segmentation, and classification on images
- Use cases: quality control, counting, shape or defect analysis
- Key information extraction from business documents
- RAG to enrich LLM responses with internal data
- Prompt engineering, validation, and evaluation of AI outputs
- Transformers and Natural Language Processing (NLP)
- Use cases: data extraction from PDFs, report summarization, support team assistance
- Data cleaning and preparation (structured, text, images...)
- Custom ML/DL model development (classification, regression, clustering...)
- Model performance optimization
- Evaluation and validation before production deployment
- AWS deployment (SageMaker, ECR, Fargate, S3) + Docker
- REST API development with FastAPI
- Automated testing, CI/CD
- Monitoring and reporting interface (Streamlit, MLflow)
- Sharing MLOps best practices with existing teams
- Simplifying AI concepts
French
Native or bilingual
English
Conversational
Experience
- Mas Seeds groupe MaïsadourData Scientist / Machine Learning EngineerRAW MATERIALS INDUSTRYSeptember 2024 - February 2026 (1 year and 5 months)Bordeaux, FranceThe challenge:From a fragmented and partially manual approach to automated, innovative, and competitive R&D using ML models and advanced analytics tools.My mission was to modernize analytical tools and leverage existing data to improve varietal selection, accelerate R&D, automate processes, and innovate in varietal creation, in line with business objectives.I focused on 5 key areas:1️⃣ Identification and optimization of more performant ML/DL models on multi-trait NIRS data to improve corn composition prediction and secure the selection of better varieties.2️⃣ Implementation of an automated PDF data extraction system (RAG + LLM) to save time and enhance R&D productivity.3️⃣ Development of a RandomForest model to classify sunflower genotypic data and ensure the creation of high-quality varieties.4️⃣ Testing and evaluation of an innovative Transformer model specialized in genomic data to identify genes of interest and improve varietal selection methods.5️⃣ Object detection and segmentation on images (corn ears and kernels), industrialized on AWS, to ensure accuracy and speed in yield and productivity calculations.Results:✅ 18% reduction in average error for NIRS models, improving the reliability of varietal selection decisions.✅ Automation of PDF extractions, generating significant productivity gains for R&D.✅ RandomForest classification and genotypic data analysis, optimizing the description of genetic diversity and the creation of new varieties.✅ Accelerated and more reliable yield and productivity calculations through Computer Vision.✅ Overall contribution to faster, more reliable, and competitive R&D, enhancing customer trust and innovation capacity against competitors.
- Openclassrooms –CentraleSupélecProfessional BreakEDUCATION AND E-LEARNINGSeptember 2022 - August 2024 (1 year and 11 months)Bordeaux, FranceTraining with Openclassrooms:
- Seven applied projects on business cases (banking, retail, energy, health) validated through presentations.
- ML & DL in Python, Production Deployment/MLOps: API, dashboards, AWS, PySpark, Computer Vision.
Python, Deep Learning, NLP, CNN, scikit-learn, TensorFlow, AWS, Streamlit, Flask - Education NationaleStrategic Leadership and Resource Management (Assistant School Principal)PUBLIC SECTORSeptember 2020 - August 2022 (1 year and 11 months)Bordeaux, FranceMy management experience allows me to approach ML projects not only from a technical perspective but also from budgetary, organizational, and political angles:1️⃣ Complex Project Management: Leading pedagogical policy and coordinating multidisciplinary teams.2️⃣ Innovation Management: Implementing digital solutions and optimizing material/budgetary resources.3️⃣ HR Management: Leading multidisciplinary teams, conflict resolution, and change management.4️⃣ Strategic Leadership:⠀⠀✅ Organization and continuity of service: HR planning, unforeseen event management, and communication oversight.⠀⠀✅ Implementation of pedagogical policy.⠀⠀✅ Budget construction and management under an objectives agreement with the relevant territorial authority, respecting 7 fundamental principles including balance and accuracy.
Reviews
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Machine Learning EngineerOpenclassrooms2026Huit projets appliqués sur des cas métier multi-disciplinaires, évalués en soutenance : - Analyse de données sur l’alimentation, - Catégorisation de données textuelles et image, - Segmentation client, - Prédiction efficacité énergétique de batiment, - Déploiement d’infrastructure Big Data dans le Cloud, - Gestion de projet IA --> Python, ML, Deep Learning, NLP, Scikit-Learn, Tensorflow, AWS, Streamlit, FastAPI
- Data ScientistOpenclassrooms-CentraleSupélec2023- Sept projets appliqués sur des cas métier (banque, retail, énergie, santé) validés en soutenance - ML & DL en Python , Mise en production/ MLOps : API, dashboards, AWS, PySpark --> Python, Deep Learning, NLP, CNN, scikit-learn, TensorFlow, AWS, Streamlit, Flask
Certifications
- Machine Learning Engineer - Expert in Data Engineering and ScienceOpenClassrooms2026
- Data Scientist - Expert in Data ScienceOpenClassrooms2023