MLOps engagement: design, deployment and industrialization of intelligent automated processing systems for complex documents in the energy sector.

• Designed and industrialized ML pipelines for automated document analysis across PDFs, images and forms.
• Designed specialized AI agents for document processing with Textract, Rekognition, ADK and AgentCore.
• Deployed, automated and monitored models on scalable cloud architectures with Lambda, Step Functions and DynamoDB.
• Implemented production alerting and supervision mechanisms to improve system robustness with CloudWatch, SNS and SQS.
• Provided technical leadership for complex automation challenges.

NeuRoam project: platform for collecting and analyzing everyday-life behavioral data.

• Full-stack development of a web application, APIs and a mobile application for NeuRoam.
• Integrated AI models to extract spatio-temporal insights from geolocated data.
• Setting up CI/CD pipelines with GitHub and Docker.
• GDPR compliance and CNRS requirements for health data management.
• Training and support for users for the adoption of the platform.

BrainWatch project: designed AI systems to interpret human behavior for a deep-tech research tax credit project.

• Secured non-dilutive funding through CIR, French Tech and BPI programs.
• Designed scalable AI infrastructure for data-driven operations.
• Developed BrainWatch, a multimodal behavioral analysis platform.
• Implemented automated CI/CD pipelines and GDPR compliance practices.
• Optimized data processing pipelines on AWS.

AGPA project: research on the neural basis of agent-versus-patient social representation.

• Deployed ML and deep learning models for fMRI and EEG data.
• Analyzed millions of voxels to study brain-behavior interactions.
• Processed complex multimodal time series.

Research project: modeling episodic memory mechanisms in virtual reality.

• Analyzed fMRI and behavioral data across millions of data points.
• Developed a mobile application for data collection.
• Built a web interface to visualize experimental results.
• Deployed hybrid cloud infrastructure with AWS and CRNL resources.
• Published three scientific articles based on the results.

HUBBLE Learn project: behavioral observatory based on e-learning traces.

• Collected and analyzed behavioral data from medical students.
• Evaluated the impact of teaching strategies on learning outcomes.
• Designed interactive dashboards and a dedicated database.

Tamagocours project: digital game to raise awareness of educational uses of digital technology.

• Analyzed serious game adoption dynamics through trace data processing.
• Clustering and ML application to segment user behavior.
• Produced recommendations to optimize user experience and educational impact.

LENA project: validation of a speech recognition system for young children.

• Processing and analysis of the oral productions of French-speaking children.
• Developed robust NLP models adapted to child-specific speech characteristics.
• Scientifically validated the system in research and therapeutic contexts.

ASTERISK project: deployment of an IPBX for CPAM telecom management.

• Deployed an IPBX solution with Asterisk for CPAM agencies.
• Optimized networking with routers, VLANs and VMware virtualization.
• Provided technical support and maintenance for network and telecom infrastructure.