Short Term Consultant - Software Engineer (Applied AI)

About the World Bank and eMBeD

The World Bank is a global development institution committed to reducing poverty and promoting shared prosperity. Through its Development Economics Vice Presidency (DEC), the Bank generates cutting-edge research and data to inform policy design and implementation worldwide. Within DEC, the Mind, Behavior, and Development Unit (eMBeD) applies insights from behavioral science to improve development outcomes. The unit works across sectors—including education, labor markets, public health, and digital services—to design, test, and scale interventions that improve decision-making and service delivery. Increasingly, eMBeD integrates artificial intelligence (AI) into its operational and research portfolio to accelerate learning, enhance service delivery, and support governments in adopting evidence-based and human-centered innovations.

Role Purpose

The Software Engineer (Applied AI) is responsible for the production engineering of eMBeD’s AI-enabled platforms and for establishing the engineering substrate that enables non-engineering team members to contribute safely and effectively at scale. The role combines hands-on systems engineering, AI application development, and institutional capacity building, ensuring that AI solutions are robust, secure, and aligned with responsible AI standards within a development context.

Duties and Accountabilities

The Software Engineer (Applied AI) will:

• Perform advanced-level engineering work on AI-enabled applications with limited supervision, including system architecture design, production deployment, and operational ownership.
• Propose and implement solutions to moderately complex technical challenges in applied AI engineering, including agentic memory design, Model Context Protocol (MCP) server architecture, and evaluation infrastructure.
• Translate engineering trade-offs into clear, outcome-oriented language for non-technical stakeholders, including unit leadership and behavioral science teams.
• Provide technical guidance, code review, and architectural input to behavioral scientists and economists working in AI-assisted development environments.
• Serve as a technical resource for security, reliability, and responsible AI considerations across the unit’s portfolio.

Key Responsibilities

1. Production Engineering for AI Applications

• Design, implement, and maintain MCP server infrastructure and reusable agentic patterns supporting eMBeD’s AI-enabled applications.
• Manage cloud infrastructure, container orchestration, CI/CD pipelines, and deployment environments across development, staging, and production.
• Ensure scalability, reliability, and performance of deployed AI applications.

2. Engineering Standards and Quality Assurance
Establish and enforce engineering standards for AI-assisted development, including:

• Code review protocols
• Automated testing frameworks
• Evaluation harnesses
• Security checklists

Review code and architectural outputs generated using agentic coding tools, maintaining quality without limiting team agility. Develop and maintain evaluation systems, including:

• Golden datasets
• LLM-as-judge pipelines with bias controls (verbosity, position, sycophancy)
• Human-in-the-loop review workflows

Document architecture, standards, and operational procedures as living knowledge assets. 3. Security, Privacy, and Responsible AI
Implement security controls across the AI application lifecycle, including authentication, role-based access control, secrets management, and encryption. Design and test safeguards against AI-specific threats such as prompt injection, data exfiltration, and model misuse. Implement guardrails, output validation mechanisms, and content filtering aligned with responsible AI principles. Ensure compliance with data governance standards, including handling of sensitive data, third-party API integrations, retention policies, and audit logging. 4. Cross-functional Collaboration and Knowledge Transfer
Collaborate closely with behavioral scientists, economists, and operational teams to translate research and policy needs into technical systems. Enable non-engineering staff to contribute safely to AI-assisted development through onboarding, tooling support, and best practices. Communicate technical decisions, risks, and trade-offs through clear documentation suited to an asynchronous, knowledge-driven environment. 5. Continuous Improvement and Innovation
Evaluate emerging AI technologies, agentic architectures, and knowledge system approaches beyond traditional retrieval-augmented generation. Identify and address technical debt, balancing rapid delivery with long-term system sustainability. Maintain awareness of the evolving AI and engineering landscape, proactively introducing tools and practices that enhance productivity and output quality.

Selection Criteria

1. Education and Experience
Master’s degree in Computer Science, Software Engineering, Information Systems, or a related field with at least 5 years of relevant experience; or Bachelor’s degree with at least 7 years of relevant experience; or equivalent combination of education and experience. Demonstrated track record of taking AI applications from prototype to production at scale. 2. Backend and Systems Engineering
Strong proficiency in Python and Node.js or TypeScript. Experience designing and operating production APIs (REST, GraphQL, MCP servers). Solid understanding of distributed systems (queues, retries, idempotency, asynchronous processing). Experience with relational and non-relational databases, including vector stores. Proven ability to convert prototypes into production-grade systems. 3. Cloud Architecture and Production Operations
Deep experience with cloud platforms (AWS, Azure, or GCP), including containerization, orchestration, autoscaling, and CI/CD pipelines. Strong observability practices (logging, metrics, distributed tracing). Experience managing production systems, including incident response, capacity planning, and cost optimization. Preference for managed, scalable infrastructure in lean team environments. 4. Applied AI and Knowledge Systems
Hands-on experience integrating LLM APIs (e.g., OpenAI, Anthropic, open-source models via vLLM). Strong expertise in context engineering: Retrieval, Agentic memory (working, episodic, semantic), Prompt caching and long-context strategies. Familiarity with hybrid search, knowledge graphs, and advanced knowledge systems. Experience designing evaluation frameworks (gold datasets, LLM-as-judge, uncertainty estimation). 5. AI Security and Responsible AI
Experience mitigating AI-specific risks (prompt injection, exfiltration, misuse). Knowledge of guardrails, structured outputs, and content filtering. Exposure to red-teaming AI systems or contributing to AI risk assessments. Familiarity with responsible AI principles, including transparency and calibration. 6. Data Governance and Privacy
Knowledge of data protection frameworks (e.g., GDPR, OECD principles). Experience handling PII, implementing data minimization strategies, and designing retention policies. Ability to assess third-party API data flows and implement audit mechanisms. 7. AI-Assisted Development and Code Review
Daily fluency with agentic coding tools (e.g., Claude Code, Cursor, Codex CLI). Strong ability to rigorously review AI-generated code. Experience working in environments where non-engineers contribute code, with a demonstrated ability to enable contributions without compromising quality. Balance between rapid iteration and disciplined engineering standards. 8. Cross-functional Collaboration
Experience working in multidisciplinary teams (research, policy, or operations). Excellent written communication skills in document-driven environments. Ability to translate technical trade-offs into outcomes-oriented language for senior stakeholders. Track record of producing technical documentation, runbooks, and decision records.

Deliverables

The Software Engineer (Applied AI) is expected to deliver:

• Production-ready AI-enabled platforms and infrastructure
• Engineering standards and documentation frameworks
• Secure, compliant, and scalable AI systems
• Evaluation pipelines and quality assurance mechanisms
• Capacity-building support for non-engineering contributors

Reporting and Collaboration

The consultant will work under the supervision of the eMBeD leadership team and collaborate closely with behavioral scientists, economists, and operational teams across the World Bank. Please submit your CV and cover letter to embed@worldbank.org by May 15, 2026.