Location: Oxford, United Kingdom or remote

Job type: Full- or part-time

Reports to: Director of Research

Overview of the role

Meedan is seeking a Research Scientist to play a key role in our research and engineering team. Meedan works on three major projects at present:

• We build Check, a web platform for collaborative media annotation and claim-checking used by some of the world’s leading fact-checking and media organizations.
• We are also building a new AI/ML software system with local newsrooms that integrates knowledge graphs and LLMs through a community GraphRAG approach.
• Finally, we have a research project aiming to improve the detection of Technology-Facilitated Gender-Based Violence (TFGBV) in Arabic, Swahili/Kiswahili, and online gaming. This involves overseeing human-in-the-loop annotation to create datasets to benchmark and improve current approaches for detecting TFGBV.

The research scientist will conduct research and evaluation, publish high-quality, peer-reviewed papers, and help build and maintain Meedan’s software and programs. Their problem-solving skills, attention to detail, ability to learn fast, sense of initiative and positive attitude will ensure a long and rewarding tenure in a similarly-minded group.

As part of Meedan you will have the opportunity to help improve the global information ecosystem for global and local audiences. This is an ideal position for someone wanting to grow with a dynamic and diverse technology organization focused on solving hard information distribution problems. We encourage diverse applicants from underrepresented communities.

Responsibilities

• Oversee human-in-the-loop annotation of possible TFGBV content as well as build, benchmark, and evaluate classifiers for TFGBV
• Write Python scripts to facilitate data collection and integrate with Check’s API
• Research knowledge graph and knowledge representation problems, supporting the engineering team in developing production-level software solutions.
• Assist with the monitoring and evaluation of Meedan’s machine learning services such as evaluation of GraphRAG pipelines.
• Develop and undertake state-of-the-art research tasks in computational social science, social data science, computational social science, and machine learning in coordination with the Director of Research.
• Writing, publication, and dissemination of results in articles, blog posts, and project reports aimed at both academic and non-academic audiences.

Qualifications

Required

• Demonstrated experience evaluating and training machine learning systems, ideally with textual or multimodal inputs
• Excellence with Python
• Solid experience with version control, preferably Git and GitHub
• Strong command of English language and communication skills, especially with remote, multicultural teams

Preferred

• Experience building API-related web technologies: HTTP, REST, etc.
• Experience with automated testing
• Familiarity with Docker or other containerization technologies
• Experience creating and querying Knowledge Graphs using LLMs
• Experience building production software on open-source platforms

About Meedan

Meedan is a global technology not-for-profit that builds software and programmatic initiatives to strengthen journalism, digital literacy, and accessibility of information online and off. We develop open-source tools for creating and sharing context on digital media through annotation, verification, archival, and translation. We work with technologists, newsrooms, NGOs and academic institutions on award-winning projects to make online ecosystems safer, more inclusive, and more equitable, so that existing pathways for content production and distribution can better serve the needs of all communities, regardless of their locations, languages, or incomes.

Equity, Diversity and Inclusion Statement

We actively recruit and encourage applications from individuals from historically underrepresented communities, including those who identify as LGBTQ+, women, people of color, and people with disabilities.