David Bertoin

David Bertoin is a Research Engineer at Photoroom, focusing on training diffusion models for image editing. He concurrently serves as a part-time Associate Professor at INSA Toulouse, where he teaches and conducts research. Previously, he worked as a Research Scientist at IRT Saint-Exupéry, contributing to the DEEL and ANITI research programs. His research interests center on representation learning, computer vision, and reinforcement learning, with a particular emphasis on improving generalization and robustness in these domains. Bertoin actively participates in academic conferences, presenting work such as Local Feature Swapping at ICLR 2022 and studies on ReLU activation functions at Neurips 2021. His GitHub repositories reflect his engagement in projects like CLOP (Local Feature Swapping) and teaching modules for M2 MAPI3. He maintains an active presence in open-source contributions related to deep learning and reinforcement learning frameworks.

Education details are not explicitly provided in the text, but his professional trajectory suggests advanced qualifications in computer science or artificial intelligence. His research outputs span theoretical advancements in neural networks and practical applications in autonomous systems, cybersecurity, and drone technology. Collaborations within ANITI and DEEL highlight his involvement in interdisciplinary AI initiatives.