Experience

• 

  Ereputationdefender

  Jul 2018 - Sept 2018

  Développeur web

  Développement Web et SEO (Search Engine Optimisation)

• 

  Air new zealand

  Mar 2020 - Jul 2020

  Data scientist

  Developed a predictive model to forecast the number of containers of various types used on each flight for Air New Zealand. • Built machine learning models to predict container usage on flights. • Integrated these predictions into a linear optimization program to determine the optimal distribution of empty containers across airports. • Ensured equitable allocation of containers, improving operational efficiency and reducing costs.

• 

  Systel-sa

  Feb 2021 - Dec 2022

  Machine learning engineer

  Worked on the development, deployment, and optimization of machine learning models, focusing on time series, NLP, and speech recognition, with a strong emphasis on productionizing solutions. • Built and deployed machine learning models for time series analysis, NLP, and speech recognition, including geographic segmentation (using sklearn and rasterio). • Developed APIs and dashboards to provide seamless access to models and insights. • Implemented geospatial operational load estimation for various zones (hourly, daily forecast). • Utilized Docker and FastAPI for model deployment. • Applied NLP techniques to analyze intervention reports, enhancing automated data extraction. • Developed a POC application using pre-trained models for real-time or near-real-time transcription, language detection, zero-shot classification, and translation. • Optimized model inference speed by 3x to 10x using ONNX and Hugging Face Optimum.Contributed to delivering efficient, scalable, and real-time solutions, significantly improving performance and operational efficiency. Show less

• 

  Alten

  Jan 2023 - now

  As part of a team focused on refactoring the backend of a web application, I played a key role in transitioning from a legacy system built with Go on AWS Lambda and API Gateway to a modern Python-based API leveraging FastAPI, hosted on AWS ECS with Fargate.Key Responsibilities: • Implemented infrastructure as code using Terraform to manage AWS services (ECS, Load Balancer, CloudFront, Cognito, RDS, S3). • Developed and maintained the CI/CD pipeline with Jenkins, automating builds, tests, and deployments of Docker containers and reducing deployment times. • Containerization and Docker Management: Managed the Docker container used to host the new application, optimizing for performance, security, and compliance. • Maintained the legacy infrastructure while building isolated development environments to support feature testing for developers.Key Achievements: • Successfully transitioned backend to a more scalable and efficient architecture. • Enabled smoother and faster deployments through the creation of a robust CI/CD pipeline. • Enhanced the security of the application • Enabled smoother feature development by setting up dedicated environments for testing. Show less Helped build an operational research platform on AWS in less than a year, centralizing solver licensing and reducing costs across teams. • Migrated infrastructure from CloudFormation to Python AWS CDK for better maintainability. • Built a CI/CD pipeline with Jenkins for seamless multi-environment deployments. • Leveraged AWS services like EC2, Fargate, and Lambda for a scalable solution. • Solved hardware-based license challenges to optimize platform usage.The platform enabled teams to offload solver jobs without individual licenses, significantly lowering operational costs. Show less Worked on the machine learning component of a data application aimed at comparing industrial parts to identify similar ones. The project, which had struggled to scale and validate results over three years, was successfully scaled and optimized through my contributions. • Optimized ML workflows to enable processing on the full dataset (previously limited to a small subset), significantly improving scalability and efficiency. • Enhanced result validation by implementing best practices, improving both the quality of the results and the performance of the model. • Collaborated with business teams to refine the ML pipeline, reducing compute costs and processing time while improving result accuracy. • Developed the solution in PySpark on AWS Glue, using a combination of NLP, data engineering, and Random Forest models to process text and numeric data.The improvements led to more reliable, cost-effective, and scalable ML outcomes, directly benefiting the business. Show less

  • Cloud Engineer at Airbus

    Jul 2024 - now

  • Cloud Engineer at Airbus

    Jan 2024 - Jul 2024

  • Data Scientist at Airbus

    Jan 2023 - Jan 2024