Cultural Data Collection and Analysis Support

Volunteering

Contract

This is a UNV contract. More about UNV contracts.

The Real-Time Translation (RTT) Project seeks to strengthen communication between United Nations peacekeeping personnel and the communities they serve. It supports the development of translation tools for languages and dialects underserved by existing solutions, including in settings with limited or no connectivity. By reducing language barriers, the project seeks to strengthen the capacity of peacekeeping personnel to identify community concerns and facilitate information exchange during patrols and other field engagements.

The project is led by the Social Innovation and Special Projects Team (SISPT) within the Policy, Evaluation and Training Division of the Department of Peace Operations. Current work focuses on translation between Juba Arabic and English in support of the United Nations Mission in South Sudan (UNMISS), alongside efforts to support communication in Cypriot Greek and Cypriot Turkish in coordination with the United Nations Peacekeeping Force in Cyprus (UNFICYP). These efforts involve developing language resources and incorporating cultural context to help the tools account for local meanings, expressions and perspectives.

Cultural datasets are an important component of this work, as effective translation requires attention to the context in which language is used. Under the guidance of SISPT, the Online Volunteer will support the development and integration of structured cultural datasets into the RTT tools. Drawing on materials identified with the team and participating missions, the assignment will contribute to ongoing efforts to improve the tools’ sensitivity to local context, supported by human review and validation.

Develop a structured cultural dataset for integration into the Real-Time Translation (RTT) tool, under the guidance of SISPT. Tasks will include:

• Process selected cultural materials into a consistent data format suitable for integration into the tool. • Assess data quality and refine the dataset through linguistic and cultural review. • Support dataset integration and evaluate the tool’s handling of culturally specific content.

The deliverable will be one documented cultural dataset, including a concise description of the methodology, integration procedures and evaluation findings.

• Advanced university degree in computer science, natural language processing or a related field. • Experience developing cultural datasets and evaluation benchmarks for language models, particularly for underrepresented languages and cultures. • Experience in data processing and annotation, multilingual language model development and evaluation, and technical documentation.

Apply now

Added 18 days ago - Updated 14 days ago - Source: unv.org