AI solutions for Climate and Environment Data

Purpose of Activity/Assignment:       

• Disasters, climate change, and environmental degradation present significant and escalating threats to the well-being of children globally. In order to protect the lives, health and well-being of children and the resilience of their communities to adapt essential social services to a changing climate, more frequent/extreme disasters and a degrading environment, comprehensive child-centered risk assessment along with data systems and policies that work across development and disaster management to identify and prioritize the most vulnerable children and services are needed. This data can be used to inform multiple purposes: disaster risk reduction and climate change adaptation, preparedness planning, and wider development agendas.
• Foundational data and evidence provide UNICEF offices with the necessary support they need to integrate sustainability, low-carbon development and climate and disaster resilience into programming, operations and advocacy. This crucial first step enables teams to optimize their approach and actions to effectively engage in sustainability and climate action. In addition, demonstrating increased capacity and readiness to fulfill this mandate is increasingly essential for securing additional finance, which is crucial for maximizing the impact of our initial investment in sustainability and climate action.
• UNICEF is seeking a dedicated, proactive consultant to conduct a feasibility study and build a proof-of-concept to strengthen UNICEF’s capacity in foundational data and analysis on sustainability and climate action as it relates to children, through the integration of artificial intelligence technologies. This will include but not limited to building and testing ML-based algorithms to address data gaps and proxy data sources, test generative AI solutions to improve accessibility and supporting existing UNICEF climate and environmental data initiatives, such as the Children’s Climate Risk Index (CCRI).

Scope of Work:

Under the guidance of the programme specialist climate, energy, environment and disaster risk reduction, the consultant will perform the following duties:

• Problem Definition and Scope
• Clearly define the problem or challenge that AI can address within the context of UNICEF’s foundational climate and environment data and analysis.
• Review current process of data analysis, methods, and data pipelines, report generation and existing documentation related to climate and environment data.
• Specify the scope of the assessment, including the relevant domains (climate, environment, energy, disaster risk reduction) and the specific data types involved.
• Stakeholder Engagement and Requirements Gathering:
• Identify key stakeholders, including UNICEF staff, technical experts, and end-users.
• Gather requirements by consulting with stakeholders to understand their needs, expectations, and desired outcomes.

Feasibility Analysis:

• Assess the feasibility of implementing AI solutions, including but not limited to machine learning and/or foundation models, geoAI,GenAI and NLP solutions within UNICEF’s operational context.
• Consider technical, organizational, and resource constraints to introduce AI solutions in climate and environment data collection, analysis, visualization and adoption.

Prototype Development and Testing:

• Develop a proof-of-concept and design/build a prototype for each identified use cases.
• Test and maintain prototypes using relevant data samples to ensure streamlined operation and to evaluate performance.
• Risk Assessment and Mitigation:
• Identify potential risks associated with AI/ML adoption, such as bias, privacy, and security.
• Propose mitigation strategies to address these risks.

Evaluation Metrics and Reporting:

• Define evaluation metrics (e.g., accuracy, precision, recall) to measure the success of the AI solutions.
• Prepare a comprehensive report summarizing findings, recommendations, and limitations.

Final Report:

• Produce final feasibility assessment report including but not limited to all chapters, mappings, models, pilots, data folder links, summarizing findings, recommendations, limitations and proposed next steps for UNICEF

Terms of Reference / Key Deliverables:

• Problem Definition and Scope 1X progress report of the below deliverables where relevant: Chapter that clearly defines the problem or challenge that AI can address within the context of UNICEF’s foundational climate and environment data and analysis Chapter that reviews the current process of data analysis, report generation and existing documentation Chapter that specifies the scope of the assessment, including the relevant domains (climate, environment, energy, disaster risk reduction) and the specific data types involved – By 28 Oct 2024
• Stakeholder Engagement and Requirements Gathering  1X progress report of the below deliverables where relevant: Chapter including a mapping of key stakeholders, including UNICEF staff, technical experts, and end-users. Chapter that summarizes requirements by consulting with stakeholders to understand their needs, expectations, and desired outcomes.  – By 25 Nov 2024
• Feasibility Analysis 1X progress report of the below deliverables where relevant: Feasibility assessment report of implementing AI solutions within UNICEF’s operational context – focusing on climate and environment data collection, analysis, visualization and adoption –  By 27 Jan 2025
• Risk Assessment and Mitigation 1X progress report of the below deliverables where relevant: Chapter that identifies potential risks associated with AI adoption, such as bias, privacy, and security Chapter of proposed mitigation strategies to address these risks – By 24 Feb 2025
• Prototype Development and Testing  1X progress report of the below deliverables where relevant: Chapter with a proof-of-concept AI model or prototype detailed. Chapter summarizing the performance testing of the solutions using relevant data samples – By 12 May 2025
• Evaluation Metrics and Reporting 1X progress report of the below deliverables where relevant: Chapter with evaluation metrics (e.g., accuracy, precision, recall) defined to measure the success of the AI solution(s) Chapter with report summarizing findings, recommendations, and limitations on metrics – By 30 June 2025
• Final report 1X Final report Final feasibility assessment report including but not limited to all chapters, mappings, models, pilots, data folder links, summarizing findings, recommendations, limitations and proposed next steps for UNICEF. – By 29 Aug 2025

Qualifications

Education:

• At least a master’s degree.
• Disciplines: Advanced university degree (master’s degree or higher) in Computer Science, Artificial Intelligence, Data Science, or related field.

Experience:

• Minimum of 5 years of progressively responsible experience in designing, testing, and implementing AI solutions including but not limited to ML/LLM models. Deep understanding of AI applications to climate data is a plus.
• Experience in data engineering with a specific focus on designing and implementing ML applications
• Experience in climate and disaster risk management is a plus

Competencies/Knowledge:

• Proficiency in Python, R or other relevant programming languages and experience working with big data environment
• Excellent coordination, communication and partnership development skills is desirable
• Strong writing and communication skills, with the ability to effectively communicate complex climate financing concepts and proposals, both in writing and verbally is required.

Source: https://jobs.unicef.org/cw/en-us/job/575312