AI-Powered Evaluation Intern

UN News; New York (NY); On-site; USD 17,000 - 26,000; Full time

Job summary

UN Women is seeking an internship to support the Independent Evaluation Service in applying AI across evaluation cycles. You will help identify AI use cases, develop reusable prompts, and pilot AI-assisted methods for data extraction and coding, all while ensuring compliance with gender-responsive evaluation practices.

The role emphasizes responsible AI use, validation against source material, and creating learning materials for future use. Strong English communication is required.

Qualifications

• University studies in a related field with data analytics experience.
• Practical experience using generative AI tools, including prompts, testing and validation.
• Enrollment in graduate studies or completion within two years of graduation.

Responsibilities

• Support development of AI techniques to conduct evaluations and extract data.
• Identify data sources and AI-assisted tasks to speed desk reviews and analysis.
• Test AI-assisted qualitative coding using prompts from evaluation matrices.
• Design analytics and AI-supported tests to improve evaluation efficiency.
• Develop gender-bias checks and a gender-responsive AI checklist.

Skills

Data analytics

Prompt development

English proficiency

Education

Bachelor's degree or higher in a related field

Enrollment in graduate program or within two years of graduation

Experience with data analytics is preferred

Tools

Microsoft Office

Co-pilot

Generative AI tools

Job description

UN Women is seeking an internship to support the Independent Evaluation Service in applying AI across evaluation cycles. You will help identify AI use cases, develop reusable prompts, and pilot AI-assisted methods for data extraction and coding, all while ensuring compliance with gender-responsive evaluation practices.

The role emphasizes responsible AI use, validation against source material, and creating learning materials for future use. Strong English communication is required.