Ilyas T.

As part of this role, I led the implementation of a robust MLOps infrastructure enabling end-to-end automation and traceability of Machine Learning processes, from training to deployment.

1. Design and implementation of a hybrid Cloud/On-premise architecture for industrializing ML workflows, ensuring performance and flexibility.

2. Setup of a comprehensive traceability system using MLflow for tracking experiments (parameters, metrics, artifacts) and ensuring result reproducibility.

3. Implementation of data versioning with DVC and LakeFS, guaranteeing traceability and governance of training datasets.

4. Automation of CI/CD pipelines with:

5. Deployment and configuration of key tools:

As a PhD student in computer vision and deep learning, I gained solid experience in developing defect detection models using infrared cameras. I have a deep understanding of image processing fundamentals, as well as computer vision algorithms for thermal images.

I am particularly skilled in deep learning-based object detection and localization methods, including the use of CNNs, transfer learning, semantic segmentation, and instance segmentation. Furthermore, I have advanced experience in Python programming and proficiency with deep learning frameworks such as TensorFlow and PyTorch.

I am also familiar with camera calibration methods, which allows me to develop high-quality defect detection models. My expertise in this field enables me to make a significant contribution to any project involving computer vision and deep learning technologies.

Integrate a robotic arm for construction.

Evaluate technical constraints.

Implement a strategy for optimizing building construction.