Data Curation for Behaviour Analysis in Stands
Available: from January 2025 until September 2026
Join our innovative team for a three-month virtual internship focused on building a foundational dataset for a cutting-edge AI model. The core mission is to enable the AI to understand user behaviour and satisfaction in event stands by combining emotions, sentiments, behaviours, and actions. Your primary task will be to source and meticulously localise “movie clips” or other open-source “clips” and/or datasets that allow the analysis of user behaviour and satisfaction in event
stands and small-to-medium retail stores.
Key Responsibilities:
• Sourcing & Scene Identification: Research and select a diverse range of clips in different types of events stands to ensure a rich, unbiased dataset. You will identify and catalogue key scenes for user behaviour and satisfaction.
• Data Localisation & Annotation: Precisely timestamp the start and end of these behavioural sequences and/or clips with the above-mentioned subjects. Yourclips (with detailed annotations*) will form the basis for training the AI, focusing on the character's observable reactions and interactions.
• Taxonomy Application (optional): Utilise a structured taxonomy to tag scenes (with relevant labels for emotional cues, behavioural patterns, and actions).
Desired Skills & Outcomes:
The candidate should be detail oriented, have a keen analytical eye for scene analysis, and have a strong interest in AI and data. By the end of this internship, you will have gained hands-on experience in the crucial data preparation pipeline for computer vision and AI, contributing directly to a project at the intersection of multimedia analysis and human-computer interaction. You will deliver a curated (and annotated*) portfolio of scenes, forming the cornerstone for the AI model to analyse user behaviour and satisfaction in event stands. The work will be integrated into the project in development by the company SPIC, Digital Studio and UAlg: AI.INSIGHTS (AI.INSIGHTS - Scientific microsite).
Upon successful completion of this virtual internship, you will be able to:
• Articulate the critical role of high-quality, curated data in the development of AI models, specifically for computer vision and human-behaviour analysis.
• Execute the end-to-end process of dataset creation, from sourcing and criteria-based selection to precise data localisation and metadata tagging.
• Evaluate the challenges of bias and diversity in AI training sets by consciously sourcing from a wide range of clips.
• Construct a professional portfolio of your annotated work, demonstrating your direct contribution to a
foundational AI project.
These outcomes are designed to provide you with tangible,
in-demand skills in the fast-growing field of AI data
operations and multimedia intelligence.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them.
