Research Engineer – Weakly Supervised Visual Anomaly Detection

About the role

The University of Technology of Compiègne (UTC) is recruiting a research engineer for the Roberval Laboratory within the Mechanical Engineering department. The role is part of the PostGenAI CAP 1.6 AI for Industry project, focusing on weakly supervised detection of complex visual anomalies for future industry applications.

Key responsibilities

• Design and execute a rigorous scientific study comparing deep‑learning models for complex visual anomaly detection.
• Define evaluation criteria, model deployment configurations, and conduct experiments.
• Draft scientific articles presenting findings and recommendations for deploying Visual Large Model (VLM) methods in industry.
• Analyse cost factors such as inference time, labeling effort, prompt formulation, and power consumption across different model types.

Required profile

• Engineering degree or Master’s degree in Industrial Engineering, Computer Engineering or related field.
• Solid foundation in Python programming and familiarity with AI architectures based on large language models (LLMs) and visual large models (VLMs).
• Knowledge of visual defect inspection in industrial contexts.

Required skills

• Python
• Understanding of LLM and VLM architectures