AI & Automation Associate - J-PAL Middle East and North Africa

Summary of the Organization:

The Abdul Latif Jameel Poverty Action Lab Middle East and North Africa (J-PAL MENA), hosted at the American University in Cairo, works to reduce poverty by ensuring that policy is informed by scientific evidence. J-PAL MENA already uses AI-assisted development to build internal tools, including systems for managing grant financial documentation, supporting staff feedback, and reviewing recordings of field surveys. It is also exploring more ambitious applications, such as Arabic-language AI systems and AI-enabled research back-checks.

Summary of the Role:

J-PAL MENA is seeking an AI & Automation Associate (part-time, hybrid from Cairo) to help build and improve AI, automation, and software systems that support research, finance, and operations. Designed for an early-career candidate (0–2 years of experience), the role involves working with colleagues across teams to understand their needs and turn promising ideas into tools that are reliable, useful, secure, and maintainable. Roughly a third of the role is dedicated to securing the applications built with AI, both by defining security requirements before development and by finding and fixing weaknesses before colleagues depend on a tool.

Responsibilities:

• Understand operational problems, ask the right questions, and determine whether software, AI, automation, or a simpler solution is appropriate.
• Rapidly prototype solutions, test them with users, and take projects from initial idea through deployment and maintenance.
• Build new tools and improve existing systems used across research, finance, and internal operations.
• Define security requirements before development starts, including what data a tool will handle, who should access it, and what could go wrong, and translate these into clear instructions for AI-assisted development.
• Review AI-generated code and configurations for vulnerabilities, including weak authentication, missing authorization checks, exposed credentials, injection risks, and unsafe database or API access rules.
• Use AI tools to run security reviews and tests on applications, then manually verify the findings and confirm the fixes work.
• Check existing internal tools for security gaps and harden them, prioritizing by the sensitivity of the data involved.
• Protect systems that involve research participant data, financial records, survey recordings, or confidential staff information, including access controls, secrets management, logging, and backups.
• Document security decisions and help establish simple, repeatable checks that every tool passes before release.
• Apply sound engineering practices so tools are dependable beyond the prototype stage, including Git and version control, reviewing and understanding AI-generated code, writing tests and documentation, and considering validation, logging, backups, dependencies, and failure modes.

Qualifications:

• A bachelor's degree; computer science, software engineering, computer engineering, cybersecurity, data science, or another relevant technical field is particularly relevant.
• 0–2 years of relevant experience; recent graduates are strongly encouraged to apply.
• Experience building software using modern AI coding tools.
• Knowledge of Python, JavaScript/TypeScript, or another general-purpose programming language.
• Familiarity with Git, SQL/databases, APIs, web applications, and basic deployment.
• Ability to debug, test, and reason about code rather than relying uncritically on AI-generated output.
• A solid foundation in cybersecurity, gained through a degree, coursework, certifications, CTFs, bug bounty work, or personal projects.
• Working understanding of common web and API vulnerabilities (such as the OWASP Top 10) and how to find and fix them in real code.
• Knowledge of authentication, authorization, and access control, including database-level permissions.
• Experience with secrets management and the safe handling of credentials and sensitive data.
• Ability to think through how a system could be misused or attacked before it is built.
• Familiarity with AI-specific security risks, such as prompt injection and data leakage through LLMs, is a strong plus.
• Ability to take an ambiguous problem and independently work toward a practical solution.
• Ability to collaborate effectively with nontechnical colleagues and to learn new tools and approaches quickly.
• Ability to balance experimentation and speed with attention to reliability, security, and maintainability.
• Ability to work comfortably in English and Arabic.
• Ability to attend in person in Cairo, Egypt.