AI Scientist - Consultant
World Bank Group
📍Remote (Global) 🕔 Full Time ⏳24/08/2025
The Development Impact Group’s AI Lab is launching a recruitment process for an AI Scientist – find out more about the Lab’s work here. They will have the opportunity to work across the different projects of the AI Lab, dependent on their experiences, interests, and project needs. The candidate hired through this process will be offered employment as soon as possible and will have the option to work remotely. To apply, please send your CV and cover letter to dime_ai@worldbank.org (with “AI Scientist – Application” in subject header) by 24 August 2025 11:59 PM EST.
Background & Objectives
The Development Impact Group is part of the World Bank’s Development Economics Vice-Presidency (DEC). The Group’s purpose is to increase the use of research evidence, in particular the use of impact evaluations (IE), in the design and implementation of public policy and to develop institutional capacity and motivation for evidence-based policymaking. The Development Impact Group’s work focuses on:
Evaluating programs at scale to answer policy-relevant questions;
Developing innovative data systems and data analytics tools for real-time decisions;
Using causal inference analysis to identify mechanisms that improve results;
Building capacity to strengthen country institutions for evidence-based policymaking.
The Group aims to overcome the challenge of identifying true cause-and-effect relations in policy programs. By linking researchers to policymakers and feeding results back into policies, we foster systematic use of evidence, which informs adoption, mid-course corrections, and scale-up of policies. With a portfolio of more than 220 impact evaluations, the Group operates across all sectors in about 60 countries spanning the globe. The IEs test a variety of interventions and mechanisms to understand why policy succeeds or fails and how to improve policy design and implementation to obtain better results.
As part of the aforementioned mission, the AI Lab is pioneering the next frontier of impact evaluation, leveraging AI and machine learning to develop and implement research, interventions, and tools that address pressing global challenges. Operating with a commitment to maintaining the highest ethical standards and rigor, particularly considering vulnerable communities, the team’s work has uncovered new, evidence-based approaches to predict food insecurity, mitigate hate speech and misinformation, advance conflict early warning systems, and address gender biases. Our mission is to leverage cutting-edge technology to advance social progress and sustainable development worldwide. The team seeks to employ recent technological advancements in ML, AI, and adjacent disciplines as an enabling force for the IE work the wider Development Impact Group conducts.
Project Details
The AI Scientist (ST Consultant) can work across all three project streams (and independent research explorations):
ZeroHungerAI: Food Security threatens hundreds of millions globally, with undernourishment rising from 580 million in 2012 to 730 million in 2022 (FAO, 2024). Climate change, population growth, political instability, and economic disparities worsen the issue, hitting vulnerable communities hardest. Predicting these crises is challenging due to outdated data. DIME AI’s research pilot (research paper, blog article) uses natural language processing on non-traditional data sources, such as news articles, to detect early signs of food crises, improving forecasts by up to 46% a year in advance, allowing more time for action. This provides extra time for proactive measures to be taken. Real-time news integration aids decisionmaking in data-scarce environments, improving humanitarian responses and saving lives. The team is now looking into turning the successful research pilot from 37 countries into a live public good with weekly updates at the district level across 82 countries. Additionally, the project will consist of research exploration, including adding local languages to improve predictive accuracy and constructing geospatial signals to provide higher-resolution predictions. The aim of the project is also to be integrated with existing food insecurity and anticipatory action programming.
ImpactAI, DIME AI’s flagship project, harnesses the power of large language models (LLMs) to sift through swathes of impact evaluation research and accurately extract insights that provide effortless access to causal evidence for improved decision-making. With an added generative AI layer, this tool lets users provide natural language queries and receive personalized, precise responses. Tailored for a diverse audience committed to global development and humanitarian efforts, primary users include development practitioners, policymakers, international development organizations, academic institutions, research institutes, think tanks, and the economic research community. As a one-stop source that lowers the barriers to accessing high-quality research, ImpactAI offers easy access to cutting-edge ideas, best practices, and empirical evidence, thereby enhancing the effectiveness of development projects worldwide. Learn more about the product here.
SocialAI: Bias, Hate Speech, and Misinformation are disproportionately and increasingly affecting marginalized groups in the Global South. Hate speech detection (HSD) models help combat harmful online content but often rely on U.S.-centric data, overlooking other English dialects and other languages worldwide. Conventional testing methods, often involving biased datasets, can also inflate content moderation performance. Culturally tailored HSD models combined with human oversight can better identify hateful content. DIME AI’s previous research (research paper) on Nigerian tweets shows that combining culturally tailored models with human review of just 1% of flagged tweets can moderate 60% of hateful content. Understanding social media usage helps track societal trends and improve policy responses, such as curbing interethnic violence. DIME AI uses its top-performing model to identify Nigerian users sharing hate content and targets them with celebrity-led prosocial ads. Building from this knowledge, DIME AI is now scaling this line of work to additional countries as well as developing an Entertainment Media Gender Monitor for conducting content analysis of programming consumed by adolescent girls and young women to focus on gender norms. Simultaneously, we will be evaluating context-specific gender bias exhibited by LLMs when operating in the Global South.
Responsibilities
The AI Scientist will coordinate their work with the project PIs and other research team members. You will collaborate closely with the domain experts and incorporate real-world constraints and validity checks into the models. Specific duties include, but are not limited to:
Conceptualize, design, and refine research prototypes to improve on existing AI/ML pipelines
Experiment with model architectures, i.e., test and compare different modeling approaches and architectures to identify the most effective solutions for achieving the project's goals in information extraction and domain adaptation.
Fine-tune (large) language models such as Llama, Gemini, or GPT to the specific requirements of the project through domain-specific fine-tuning on the prepared corpora.
Develop a robust evaluation framework (incl. statistical diagnostics, out-of-sample testing, and scalability), ensuring alignment with project objectives.
Collaborate with the AI/ML Engineers to ensure smooth handoff of models to production environments (incl. packaging and deploying trained models in cloud infrastructure, ensuring scalability and low-latency performance for end-users).
Target publication in top-tier NLP and AI conferences by documenting methodologies, experimental results, and practical insights gained throughout the development cycle.
Stay current on research trends in AI, machine learning, and relevant domain knowledge.
Communicate findings and model results to stakeholders, ensuring interpretability for policy and development contexts
Requirements
3 years of hands-on experience in an AI/NLP/ML role focusing on methods such as encoder models, decoder models, knowledge graphs, and AI-based approaches to time series forecasting
Advanced degree (PhD preferred or MSc) in Machine Learning, Computer Science, Statistics, or a related field
Strong proficiency in Python, with experience in ML frameworks (PyTorch, TensorFlow, scikit-learn)
Experience with large language models (incl. training them) and traditional machine learning methods
Strong statistical analysis and experimental design skills
Excellent analytical and communication skills to convey findings to non-technical audiences
Publications or conference presentations on AI and machine learning
Motivated, self-starter, able to work with little supervision, with demonstrated ability to work effectively and sensitively in teams
Preferred Qualifications
Proven track record of handling complex, multivariate datasets and performing causal inference or forecasting studies
Experience working with humanitarian organizations or development agencies
Familiarity with big data ecosystems and parallelized computing methods
Contract Arrangements
The AI Scientist will be hired as a short-term consultant (STC) starting as soon as possible. They will be able to work remotely or in person in Washington, DC. Onboarding will include access to internal World Bank systems and technical support as needed.
Intellectual Property
The World Bank shall, solely and exclusively, own all rights in and to any work created in connection with this agreement, including all data, documents, information, copyrights, patents, trademarks, trade secrets or other proprietary rights in and to the work. The AI Scientist is not allowed to post or publish (electronically or in print) any project-related information without the explicit permission of the Impact Evaluation Team
Apply Now
To apply, please send your CV and cover letter to dime_ai@worldbank.org by 24 August 2025, 11:59 PM EST. Only shortlisted candidates will be contacted. We are proud to be an equal opportunity and inclusive employer with a dedicated and committed workforce, and do not discriminate based on gender, gender identity, religion, race, ethnicity, sexual orientation, or disability. Learn more about working at the World Bank and IFC, including our values and inspiring stories.
