Artificial Intelligence Engineer/Researcher - ABM/LLM Integration

The overall objective of this consultancy is to research, develop, and validate an integrated ABM+LLM framework capable of supporting evidence-based, AI-enhanced policy analysis for the Arab region. Specific objectives include: - Deliver a validated, calibrated rule-based ABM model and document its policy simulation capabilities; - Conduct a systematic literature review establishing the state of the art in ABM+LLM integration; - Design and implement an ABM+LLM architecture with LLMs functioning as tools and as autonomous agents; - Benchmark the integrated model against alternative LLM-based decision-support approaches;

Background: The United Nations Economic and Social Commission for Western Asia (ESCWA) is mandated to support the sustainable development of the Arab region through evidence-based policymaking, data analytics, and digital innovation. The Decision-Support and Data Science Division (DSDS) leads ESCWA's efforts in developing advanced analytical tools and AI-powered platforms to assist member states in formulating effective public policies. Agent-Based Modelling (ABM) constitutes one of the methodologies used for simulating complex socioeconomic systems, enabling policymakers to assess the potential impact of interventions in non-linear, dynamic environments. The rapid advancement of Large Language Models (LLMs) presents a transformative opportunity to augment ABM frameworks with natural language reasoning, contextual intelligence, and adaptive policy advisory capabilities. In this context, DSDS seeks to hire a junior AI Engineer/Researcher to undertake a rigorous, phased research and development initiative integrating LLMs into ABM architectures. The outputs of this initiative will directly contribute to ESCWA's AI for Policy program and will yield both publications and deployable decision-support prototypes for use by ESCWA member states. Duties and responsibilities: Under the overall supervision of the Chief of the Decision-Support and Data Science Division (DSDS), and in close coordination with the designated ESCWA focal point, the Consultant shall carry out the following tasks: Phase 1: ABM Calibration and Validation (June – July 2026) - Improve calibration performance and robustness of the existing ABM model against current data targets for the rule-based ABM model. - Run comprehensive scalability tests across number of agents, simulation steps, calibration iterations, and simulation runs. - Define a historical policy test set, target data, and validation metrics in alignment with DSDS standards. - Simulate emerging policy scenarios, analyze results, and produce policy-relevant insights. - Prepare and submit a draft report on the findings Phase 2: Literature Review and Architecture Design (August – September 2026) - Conduct a comprehensive and systematic literature review covering the ABM+LLM research landscape, including key methodologies, benchmarks, and applications. - Define the conceptual architecture for ABM+LLM integration, specifying the roles of LLMs as contextualizers and policy advisors embedded within the ABM framework. - Document the architecture, methodology, and a concrete case study relevant to Arab regional development contexts as assigned by Chief DSDS Phase 3: LLM Integration and Benchmarking (October – November 2026) - Develop and deploy the first version of the LLM-enhanced ABM model, integrating LLMs as analytical tools within the simulation environment. - Conduct rigorous benchmarking of the LLM-enhanced ABM against established LLM-based decision-support approaches, including standard LLM, RAG-based LLM, and MAS LLM configurations. - Integrate LLMs as autonomous agents and share the model with assigned domain experts for participatory validation and human-in-the-loop refinement. - Deploy the first version of the LLM-empowered ABM model to the DSDS GitHub repository.

A Master's in Computer Science, Artificial Intelligence, Computational Social Science, Applied Mathematics, Physics, or a closely related discipline is required. A first-level university degree with a minimum of three (3) years of relevant experience may be accepted in lieu of an advanced degree. All candidates must submit a copy of the required educational degree. Incomplete applications will not be reviewed. A minimum of two (2) years of demonstrated experience in agent-based modelling, complex systems simulation, or computational social science is required. Proven experience with large language models, including prompt engineering, fine-tuning, and integration with external systems or APIs is required. Experience with Python-based ABM frameworks and LLM integration libraries is required. Familiarity with policy-relevant simulation use cases, preferably in socioeconomic and/or development contexts is desirable. Demonstrated proficiency with version control (Git/GitHub) and reproducible research practices is required. Track record of academic publication in peer-reviewed journals or major AI/computational science conferences is desirable. Strong analytical and research skills with the ability to synthesize complex technical and policy information is required. Excellent written communication skills in English; ability to produce high-quality technical reports and papers is required. Capacity to work independently, meet deadlines, and deliver high-quality outputs with minimal supervision is required. Commitment to the values and principles of the United Nations is required.

English and French are the working languages of the United Nations Secretariat; and Arabic is a working language of ESCWA. For this position, fluency in English is required. Note: “Fluency” equals a rating of ‘fluent’ in all four areas (speak, read, write, and understand) and “Knowledge of” equals a rating of ‘confident’ in two of the four areas.

Not available.