WBG Pioneer - AI Engineer Intern

Description du poste:

WBG Pioneer Background/Organizational Context The Integrity Vice Presidency (INT) investigates and helps prevent fraud and corruption in World Bank Group-financed projects. Within INT, the Data Lab is a multidisciplinary team spanning data, AI, analytics, and cloud engineering that builds and operates production-grade, responsible AI and data solutions to enable and streamline INT's core business processes. This 150-day internship offers a hands-on opportunity to work at the frontier of enterprise agentic AI. Under the direct supervision and mentorship of the Data Lab Team Lead, the AI Engineer Intern will learn to design, develop, and deploy AI agents that support INT's core functions. The role is well suited to a fast learner who stays aligned with industry trends in enterprise agentic AI and is eager to apply modern AI techniques to real-world integrity use cases within a secure, governed cloud environment. Duties and Responsibilities * Learn and apply the Data Lab's AI development practices to help design, build, and fine-tune AI agents and LLM-based solutions that streamline INT business processes. * Develop, test, and iterate on prompts to improve the accuracy, relevance, and reliability of AI outputs. * Support the optimization of Retrieval-Augmented Generation (RAG) pipelines-indexing, retrieval, and inference-for INT's AI agents. * Use Azure cloud resources for data storage, processing, and retrieval to support AI model requirements. * Collaborate with front-end, back-end, and DevOps team members to integrate AI models into applications and resolve integration issues. * Explore and evaluate emerging enterprise agentic AI frameworks and tools, sharing findings and recommendations with the team. * Maintain clear, up-to-date documentation of AI models, prompts, design decisions, and change logs to ensure transparency and traceability. * Participate in team meetings, provide regular progress updates, and present work to the Data Lab Team Lead and management