Samy A. Foudil

I design and industrialize document automation workflows that extract, enrich, and interpret complex PDFs, images, and forms; the goal is to make unstructured information capture reliable for business units.

- ML & LLM Pipelines: SageMaker orchestration of Textract, Rekognition, and LLM models (Bedrock / open-source) to segment, contextualize, and structure data.

- Scalable cloud architecture: serverless deployment (Lambda, Step Functions, EventBridge) with CloudWatch / Prometheus monitoring, real-time alerting, and Terraform CI/CD.

- Production MLOps: automation of training, model registry, blue-green deployments, and A/B testing to optimize accuracy and latency.

- Security & compliance: KMS encryption, IAM least-privilege, GDPR anonymization, traceable logging..

Development of an innovative platform for collecting and analyzing geolocated and contextualized behavioral data, integrating AI models for extracting relevant insights. The project aims to optimize data processing and exploitation for advanced analysis of everyday behavioral data.

- Consulting and expertise in developing web applications and APIs for data collection and exploitation, integrating scalable and secure architectures. (AWS & OVH Cloud)

- Full-stack development of a web application, APIs, and a mobile application (NeuRoam - BREAL v3), optimizing real-time data management and analysis. (Django, Ninja, React Native)

- Conducted a comparative study to identify the most suitable AI modeling for analyzing collected data, applying advanced methodologies in cognitive neuroscience. (OpenAI, Mistral, Meta)

- Implementation and automation of CI/CD pipelines with GitHub and Docker, ensuring efficient and reproducible deployment of models and applications. (GitHub, Docker, OVH Load Balancer, Debian)

As CTO of the Deep Tech startup Experitem and assisted by a team of developers, I managed the development of the company's applications and infrastructure. Additionally, I led the BrainWatch research project, which focuses on developing fundamental knowledge in applied cognitive neuroscience, with the goal of developing a set of innovative methodologies for acquiring the experience of an individual and a population in their daily environment.

Implementation of the following solutions:

- Implementation of data collection and analysis routines: AWS CloudWatch + QuickSight

- Creation and maintenance of networks: VPN + AWS VPC

- AWS Cloud Infrastructure: EBS + ECS

- CI/CD pipeline on GitLab and with CodeBuild, CodePipeline

- Web application with VueJS front-end and Spring Java API

- R&D application with Django API and geolocated React Native mobile application

- GenAI with ChatBot fine-tuned on our business data and llama langchain

- Enterprise collaboration platform with Nextcloud and Odoo on-site.