People for AI

Main Features & Services

People for AI handles both complex images and text labeling projects up to thousands of annotation tasks per week. Services include defining the right data annotation or tagging strategy, selecting the best process, tool, and team.

Core Advantages & Features

• No Crowdsourcing: Works only with in-house labelers to ensure world-class labeled data and a higher security level.
• Long-term Labelers: Hires labelers only with long-term contracts to effectively train them and manage complex, long-term projects.
• No Black-box Effect: Communication is the key; provides metrics and progression status on a regular basis.
• Experts at Hand: A deep understanding of the field is available to define data labeling strategies together.

Typical Use Cases

• Microscopy: Labeling mineral and biological data on microscope images.
• Autonomous Cars: Classification of people, traffic signs, segmentation, etc.
• Infrastructure: Identification of defects on railroads and energy networks.
• Food & Retail: Identification of foods, retail products, and precise segmentation.
• Other Industries: Autonomous vehicles, OCR and automated processing of administrative documents, content moderation on social networks, maritime, satellite and aerial images, agriculture and food, legal NLP, dermatology, etc.

Target Users

Algorithm teams and enterprises requiring machine learning training data (e.g., Foodvisor, Kili, Milvue, Alteia, Pointly, Inarix).

Pricing Information

For production projects requiring more than 500 hours of annotation, the cost is usually between €6 and €9 per annotation hour, covering 80% of projects. The price depends on the complexity for annotators and the amount of review required. This price includes annotation, review, and customer care (communication, iterative methodology, annotation instructions refinement). It does not include the selection/training of the annotation team and the setup of the annotation tool (€200-€300, depending on complexity). Smaller projects (POC) require a minimum of 50 to 100 hours of annotation.

Usage Process

Get in touch via the contact page, get assigned a project manager, build an annotation team specialized to your needs, and start the project after defining the tool and initial annotation instructions.

FAQ

• Labeling Tool: Can adapt to any kind of data labeling tool: open-source, proprietary, in-house, etc.
• Project Complexity: Labels from the most complex to the simplest projects; expert labelers are not priced the same as novice labelers.
• Quality Commitments: Committed to respecting quality KPIs set jointly with the client; re-annotation is free of charge up to 2 iterations if quality is poor.
• Privacy & Compliance: Works for large French companies on highly confidential missions and is fully GDPR-compliant.
• Team Size: As of Q3 2023, there are 6 employees in France and 80 employees in Madagascar.