1. Background and Rationale

The Centre on Armed Groups and its partner, New York University’s Center on International Cooperation (CIC), are leading a project aimed at strengthening countries’ capacity to identify, monitor, and analyse risk and protective factors for conflict prevention through the application of data-driven and artificial intelligence-enhanced (AI) tools. 

2. Objective of the Consultancy

Building on CIC’s existing online conflict and violence tool – Preventing Armed Violence An Evidence-Based Tool for Smarter Policy & Action (Prototype) –  the overall aim of the consultancy will be to expand and refine the tool’s user-friendliness and practicality by integrating an AI-component capable of identifying and weighting factors across diverse country contexts. More specifically, the consultancy is to expected to design, implement, and document an AI-assisted process to:

• Strengthen the existing database through integration of vector-based embeddings.

• Develop a replicable methodology to identify, weigh, and monitor country-specific risk and protective factors; and

• Deliver a functional docker container encapsulating all relevant models, scripts, and dependencies for deployment within the project environment.
  
  

3. Scope of Work and Activities

The consultancy will consist of three main components, as outlined below.

Component 1 – Research: Add articles to the database and train AI to identify risk and protective factors (12 days)

Activities:

• Create vector embeddings for the existing Risk/Protective Factor Database (2 days)

• Gather sample data from Ghanaian sources (5 days) 

• Tag collected documents and aggregate results (2 days) 

• Summarise identified risk and protective factors (3 days)

Key Deliverables:

1. Updated risk/protective factor database with vector embeddings.

2. Tagged dataset from Ghanaian sources.

3. Short summary report outlining identified risk and protective factors.

Component 2 – Operationalise: Identify and monitor country-relevant risk and protective factors (39 days)

Activities:

• Gather and analyse in-country data related to risk and protective factors (17 days)

• Conduct an exploratory analysis to develop methodology for weighting factors based on available data (12 days)

• Test and refine the methodology using data from the selected country (10 days)

Key Deliverables:

1. Country-specific analytical dataset.

2. Documented methodology for weighting risk and protective factors.

3. Brief report summarising test results and recommendations.

Component 3 – Wrap the application in a docker container for deployment to CIC website

Activities:

• Develop containerization strategy and environment setup

• Explore runtime and model Packaging (Ollama for self contained, OpenAI for remote - API based)

• Add/configure vector DB as a service in the Docker ecosystem (Qdrant)

• Package the App layers (React + .NET) into production containers with a unified build/deploy process

• Prepare documentation for distribution packaging

• Ensure the packaged containers run reliably across OS environments

Key Deliverables:

1. Fully functional docker container containing all developed tools and models.

2. User and technical documentation for deployment and replication.

4. Deliverables and Timeline

The expected timeline for the completion of deliverables is as follows: 

5. Qualifications and Experience

The consultant or firm should demonstrate:

• Be familiar with CIC’s Preventing Armed Violence An Evidence-Based Tool for Smarter Policy & Action (Prototype) tool or similar tools

• Proven experience in AI, machine learning, and data science applications in social science or policy research;
  
  

• Knowledge of natural language processing (NLP) and vector embedding techniques;
  
  

• Experience developing and deploying Dockerized applications;
  
  

• Strong analytical, documentation, and communication skills;
  
  

• Experience with governance, conflict, or development-related data preferred.

• Experience supporting governments in collecting and analyzing data to enhance security

6. Reporting and Supervision

The consultant will report to CIC’s technical lead. Regular coordination meetings will be scheduled to review progress, validate outputs, and ensure alignment with project objectives.

7. Duration and Payment

Payment will be made upon submission and acceptance of all deliverables, in accordance with the Centre’s procedures.