Governed cultural knowledge graph for multicultural AI - ON-1221

Genre de projet: Recherche
Discipline(s) souhaitée(s): Informatique, Sciences mathématiques, Cultural studies, Sciences sociales et humaines, Langues et linguistique
Entreprise: Tulong Technologies Inc.
Durée du projet: 4 à 6 mois
Date souhaitée de début: Dès que possible
Langue exigée: Anglais
Emplacement(s): Toronto, ON, Canada
Nombre de postes: 1
Niveau de scolarité désiré: MaîtriseDoctoratRecherche postdoctorale
Ouvert aux candidatures de personnes inscrites à un établissement à l’extérieur du Canada: No

Au sujet de l’entreprise: 

Tulong Technologies is an AI company building a plug-and-play cultural intelligence engine for enterprises and public-sector organizations. Tulong helps teams reduce and prevent AI bias by combining governed multicultural data with structured Cultural Intelligence (CQ) reasoning, confidence-aware scoring and escalation and human-in-the-loop expert review. Designed to integrate into existing AI workflows, the platform provides clear explanations, policy controls and audit-ready logs to support safer multilingual and multicultural communications and decision support across functions and industries.

Veuillez décrire le projet.: 

The primary goal of this project is to design and build a governed, culturally-annotated knowledge graph representing relationships between 50+ cultural communities for use within Tulong's Cultural Intelligence Engine (CIE).
Modern AI systems fail to represent cultural knowledge in structured, queryable form. They treat cultural context as metadata rather than as a relational system of communities, values, communication norms, diaspora connections, generational distinctions and cross-cultural relationships. This gap produces AI outputs that are culturally narrow, biased or inappropriate at scale.
The intern will design a cultural schema and taxonomy framework capturing entities such as cultural communities, subgroups, religious subdivisions, geographic variants and diaspora relationships. Using Apache AGE (a graph extension for PostgreSQL), the intern will implement a knowledge graph encoding these relationships in a machine-queryable structure. The project includes constructing four data processing pipelines covering normalization, privacy and IP governance, cultural annotation and knowledge synthesis.
The intern will develop a cultural annotation quality methodology measuring consistency, inter-annotator agreement and coverage completeness across cultural communities. The methodology will be validated against Tulong's cultural advisory council. The final deliverables include a populated knowledge graph covering 50+ cultural communities, a documented annotation framework and an academic paper on cultural knowledge graph construction methodology.
This research directly enables Layer 1 of the CIE, the governed data foundation on which all downstream AI cultural reasoning depends. Without a structured, validated knowledge graph, the CIE reasoning engine cannot retrieve accurate cultural context or detect culturally-specific assumptions in AI-generated outputs.

Expertise ou compétences exigées: 

Required: Knowledge graph design and implementation (Apache AGE, Neo4j or equivalent); PostgreSQL database architecture; Python programming for data pipeline development; ontology design and semantic web technologies (RDF, OWL); data quality measurement and inter-annotator agreement metrics (Cohen's Kappa).
Preferred: Natural language processing for cultural text annotation; familiarity with cultural studies or anthropology; understanding of data governance, privacy law and IP management; experience with multilingual datasets.
Software: Python (pandas, NumPy, NetworkX), PostgreSQL with Apache AGE, Git/GitHub, cloud compute (AWS/GCP), Jupyter notebooks.
Assets: Familiarity with Indigenous data sovereignty principles (OCAP); prior work with cultural or ethnographic datasets; experience with semantic knowledge representation.