Location: Paris (France)
Duration: 6 months (flexible start date)
Department: Fashion Team (AutoRetouch)
Internship Subject: Building Metrics for Realism and Artifact Detection in AI-Generated Fashion Models

We are seeking a motivated and detail-oriented Generative AI Intern to join our Fashion team at Meero. This internship offers an exciting opportunity to contribute to cutting-edge projects focused on improving the realism and quality of AI-generated fashion model imagery. You will work alongside a team of AI researchers and engineers on the critical task of developing quantitative metrics to evaluate and improve generative model outputs.

About the Project
Generative models are revolutionizing fashion imagery by replacing traditional photo shoots with AI-generated visuals. However, achieving realistic faces, bodies, and clothing remains challenging due to artifacts such as misaligned facial features, incorrect anatomy, and other visual inconsistencies. This internship aims to develop anomaly detection methods and quantitative image realism metrics specifically for human features of generated fashion models. Ultimately, this will enhance the quality and scalability of AI-generated fashion images.

Key Responsibilities

1. Literature Review: Conduct an in-depth review of existing metrics for assessing image realism and quality, identifying limitations specific to fashion model imagery.
2. Metric Development: Design or adapt metrics for evaluating facial consistency, anatomical correctness, and overall image coherence in generative outputs.
3. Artifact Detection Tools: Develop tools to detect common generation errors, such as extra fingers, unrealistic facial expressions, non-round pupils, or misaligned features, providing actionable insights for improvement.
4. Testing and Validation: Validate the developed metrics and tools using a diverse dataset created with Meero’s cutting-edge models, and compare results with human evaluations to ensure reliability and precision.
5. Collaborative Problem Solving: Work closely with the team to tackle anomalies iteratively, focusing on the most critical categories first to achieve meaningful and scalable results.

Of course, this internship is a collaborative effort, and you won’t be expected to address all possible anomalies at once. Our team is here to guide and support you throughout the process!

Desired Qualifications

• Educational Background: Enrolled in a Master’s or advanced undergraduate program in Computer Science, Machine Learning, Data Science, or a related field.
• Programming Skills: Proficiency in Python and experience with machine learning frameworks like PyTorch or TensorFlow.
• Experience with Computer Vision: Understanding of image processing techniques and tools for segmentation and anomaly detection.
• Analytical Skills: Strong problem-solving abilities and attention to detail.
• Team Collaboration: Excellent communication and teamwork skills, with the ability to work in a dynamic environment.
• Languages: Strong English skills (French is a plus).

Preferred Skills

• Experience with large datasets, including preprocessing and augmentation.
• Knowledge of software development best practices, such as Git, Docker, and CI/CD workflows.
• Contributions to open-source projects in generative AI or related fields.

Why Join Us?

• Gain hands-on experience working with cutting-edge generative AI and computer vision technologies.
• Collaborate with a passionate and innovative team at the forefront of AI-driven fashion solutions.
• Access to industry-leading training datasets and state-of-the-art computing resources.
• Opportunity for mentorship and professional development.

Hiring Process

1. Initial Interview: A 30-minute call with a member of our Talent Acquisition Team to discuss your background and interests.
2. Technical Interview: A 1-hour session for which you prepared a 15 min presentation of a preselected scientific article, followed by an evaluation on Python fundamentals, Machine Learning and generative AI concepts.
3. Final Interview: An on-site meeting to discuss your fit within the team and company culture.

Even if you don’t meet all the qualifications, we encourage you to apply!