Machine Learning Consultant – Req #576076

Overview

UNICEF works in some of the world’s toughest places, to reach the world’s most disadvantaged children. To save their lives. To defend their rights. To help them fulfill their potential.

Across 190 countries and territories, we work for every child, everywhere, every day, to build a better world for everyone.

And we never give up.

For every child, HOPE

Consultancy Title: Machine Learning Consultant

Section/Division/Duty Station: Division of Analysis Planning and Monitoring NYHQ

Duration: 1 November 2024 to 15 October 2025

Home/ Office Based: NYHQ / Remote

About UNICEF – DO NOT EDIT

If you are a committed, creative professional and are passionate about making a lasting difference for children, the world’s leading children’s rights organization would like to hear from you. For 70 years, UNICEF has been working on the ground in 190 countries and territories to promote children’s survival, protection and development. The world’s largest provider of vaccines for developing countries, UNICEF supports child health and nutrition, good water and sanitation, quality basic education for all boys and girls, and the protection of children from violence, exploitation, and AIDS. UNICEF is funded entirely by the voluntary contributions of individuals, businesses, foundations and governments. UNICEF has over 12,000 staff in more than 145 countries.

BACKGROUND – PLEASE REFER TO THE ATTACHMENT FOR FULL DETAILS AND INFOMATION

Purpose Of Activity/ Assignment

UNICEF requires the services of an external consultant to design, develop, test and deploy machine learning models to enhance the efficiency and effectiveness of vaccine stock management systems, subnational distribution strategies, planning processes, track adherence to stock management protocols, and strengthening vaccine accountability standard operating procedures (SOPs) through the active use of data for planning and decision-making.

The consultant will support UNICEF’s Program Group Immunization Division (PG-I) in optimizing vaccine supply chains, ensuring that life-saving vaccines reach every child, regardless of geographical or logistical challenges.

The consultant will liaise with the Data, Analytics, Planning and Monitoring Division (DAPM) and with Program Group Immunization (PG-I) to deploy robust machine-learning models within the UNICEF Azure cloud environment. These models will analyze complex data sets, process natural language protocols, optimize subnational distribution routes and schedules, and predict stock risky situations across supply chains, thereby minimizing waste and ensuring timely delivery. By integrating innovative data-driven solutions into country office’s supply chain operations, this assignment aims to strengthen the organization’s ability to deliver vaccines more efficiently and equitably, ultimately contributing to improved immunization coverage and the health of children worldwide.

Scope of Work: PLEASE REFER TO THE ATTACHMENT FOR FULL DETAILS AND INFOMATION

Plan on how to use the existing data sources together

Data sources available include vaccine and devices stocks (opening and closing balances), vaccine movement (new arrivals, distribution, wastage), forecasted demand (including at subnational levels), allocations, consumption (actual, forecasted, mean and target population), min/safety and max quantities of vaccines by supply chain level and by cold store, pipelines, causes and drivers of stockouts, country mitigation measures, outcomes of country engagement activities, coverage rates, number of zero-dose children, frequency and duration of stockouts, RTM Maturity model data, Cold Chain Inventory Data, EVM & Improvement Plans, Carbon Footprint data for the immunization programme, Waste Management Assessment Data and health Facility solarization Data. There is a need for an initial plan on how we use all the data sources together to strengthen supply chain planning and prediction models.

Develop, test, pilot and deploy robust machine learning (ML) models

Collect, clean, assess, preprocess SC data, and train ML model(s)

Terms Of Reference / Key Deliverables

Work Assignments Overview Deliverables/Outputs Delivery deadline

Coordinate with stakeholders (PG-I, DAPM) to review available datasets, create a plan on how to use all the available data sources together, what kind of ML models should be developed and why
• Development plan with the strategy to use the data sources, the ML models to be develop, and the ML techniques that will be used – By November 15 2024

Version 1. Develop, test and deploy ML models according to the development plan.

Data gaps, risks, limitations, assumptions, and potential solutions identified (Based on the findings of desk review, highlight data gaps, major risks, limitations and assumptions and propose how these would be resolved to ensure the ML models are robust, fit-for-purpose and highly accurate).

Suitability of Synthetic Data (SD) use assessed, sources /tools identified, and cost estimates prepared. (Evaluate the viability of synthetic data for model training in the absence of real-world datasets or in cases where the quality of data is ambiguous. This also includes the identification of the most cost-effective and practical SD data generation tools, cost estimates and value-for-money assessment).

Implementation plan developed and SME consulted (If synthetic data use is deemed useful, prepare a plan of action on the involvement of subject matter experts (SME), evaluation criteria, validation tests and independent third-party verification arrangements).

iSC ML model(s) developed, thoroughly trained, optimized, tested and piloted
• Version 1 of ML models developed, deployed on UNICEF Azure, and tested – By December 30 2024

Integrate the ML models with Thrive360 and the Data Control Towers

Incremental learning channeled back to improve the model(s) (The model(s) learns from its past accuracy rates and leverages the findings to improve future predictions).

Future-proofed (The model(s) follow industry standards and are flexible to upgrades, stress tests and transitions to new teams. At a minimum, the model(s) must follow best practices in modularization, consistency and versioning).
• ML models v1 integrated with Thrive360 and the Data Control Towers – By February 12 2025

Produce a document of the code and the models. All phases of model development including data source assessment, data collection, preprocessing, classification, labeling, training, model architecture, evaluation criteria, accuracy rates and results must be documented.
• Documentation produced. – March 7 2025

Identify areas of improvement and criticalities. Discuss with the stakeholders to gather feedback and additional requirements. Create an action plan for an upgrade of the models or the generation of new models
• Action plan produced – April 8 2025

Version 2. Review and improve ML models according to the action plan.
• Version 2 of ML models developed, deployed on UNICEF Azure, and tested – June 13 2025

Integrate version 2 of the ML models with Thrive360 and the Data Control Towers
• ML models v2 integrated with Thrive360 and the Data Control Towers – July 10 2025

Use the models to produce outputs for managers and country offices:

Predictive analytics developed. iSC ML models enable managers at UNICEF and partners to have access to curated predictive analytics dashboards that predict, alert and prevent stockouts, excess stocks, wastages, and other inefficiencies in stock management

Demand forecasting: UNICEF country offices and NLWGs can simulate and improve demand forecasting, especially at subnational levels

Dynamic inventory management: Managers can project stock coverage times by antigen, supply chain levels and store using single, multiple or mixed methods (push-pull, target population, vaccination session, consumption patterns or others).

Agile distribution planning: Managers are able to adjust push-pull or min-max-reorder protocols based on historical stock performance.

Inferential analysis: Managers can simulate future scenarios based on stock performance. For example, they can assess how improvement in stock management could impact coverage rates and zero-dose and vice-versa

Cold Chain Monitoring: ML model(s) are trained to enable predictive cold chain maintenance

DRIVE route optimization: The model(s) leverage DRIVE data to optimize vaccine delivery routes, frequency and quantities considering constraints such as vehicle capacity, road conditions, cold chain capacity and delivery deadlines. The algorithms must focus on reducing travel time and costs.

EVM Assessments: The model(s) must enable managers to find correlations between Effective Vaccine Management (EVM) assessments and stock performance

Overall supply chain analytics: The model(s) can produce holistic supply chain analytics. Identify inequities, inefficiencies, bottlenecks and areas for improvement, especially at subnational levels. The model should also be able to assess impacts on product/formulation changes and new vaccine introductions to the supply chain.

Management matrices: Managers have access to personalized action boards with two to three localized action points to resolve pressing operational challenges in their geographic areas.
• Outputs produced. – By September 6 2025

Update the documentation, including version 2 of the ML models, all the updated requirements and data sources, a user guide to train the models and to use the models, integration with Thrive360 and the Data Control Towers
• Documentation updated. – 15 October 2025

Please find attached TOR TERMS OF REFERENCE – MACHINE LEARNING CONSULTANT.docx

Qualifications

Education: BACHELORS

Data Science, Computer Science, Applied Mathematics, and related fields

Work Experience

Knowledge/Expertise/Skills required *:
• Proven experience in developing, training, and deploying machine learning models, particularly for supply chain optimization or similar logistical challenges.
• Expertise in supervised, unsupervised, and reinforcement learning techniques.
• Proficiency in Python.
• Experience with machine learning frameworks such as TensorFlow, PyTorch, or Scikit-learn.
• Strong experience with deploying machine learning models in cloud environments, particularly Microsoft Azure.
• Familiarity with Azure Machine Learning, Azure Functions, Azure Databricks, or similar services.
• Expertise in working with large datasets, including data cleaning, feature engineering, and data visualization.
• Experience with SQL and NoSQL databases.
• Excellent problem-solving and troubleshooting skills.
• Strong communication and collaboration skills to work effectively with cross-functional teams.
• Previous experience in similar consulting roles, particularly in global health or public health contexts, is highly desirable
• Good-to-have Skills
• Previous experience building robust data pipelines using queueing and stream processing, asynchronous patterns, parallelization, e.g., with PySpark, etc
• Understanding of vaccine supply chains, including distribution strategies, cold chain logistics, and challenges faced at subnational levels
• Ability to translate complex supply chain problems into data-driven models and solutions
• Previous experience in similar consulting roles, particularly in global health or public health contexts, is highly desirable

Requirements

Completed profile in UNICEF’s e-Recruitment system and
• Upload copy of academic credentials
• Financial proposal that will include/ reflect :
• * the costs per each deliverable and the total lump-sum for the whole assignment (in US$) to undertake the terms of reference.
• travel costs and daily subsistence allowance, if internationally recruited or travel is required as per TOR.
• Any other estimated costs: visa, health insurance, and living costs as applicable.
• Indicate your availability
• Any emergent / unforeseen duty travel and related expenses will be covered by UNICEF.
• At the time the contract is awarded, the selected candidate must have in place current health insurance coverage.
• Payment of professional fees will be based on submission of agreed satisfactory deliverables. UNICEF reserves the right to withhold payment in case the deliverables submitted are not up to the required standard or in case of delays in submitting the deliverables on the part of the consultant.

U.S. Visa Information

With the exception of the US Citizens, G4 Visa and Green Card holders, should the selected candidate and his/her household members reside in the United States under a different visa, the consultant and his/her household members are required to change their visa status to G4, and the consultant’s household members (spouse) will require an Employment Authorization Card (EAD) to be able to work, even if he/she was authorized to work under the visa held prior to switching to G4.

Only shortlisted candidates will be contacted and advance to the next stage of the selection process

For every Child, you demonstrate…

UNICEF’s core values of Commitment, Diversity and Integrity and core competencies in Communication, Working with People and Drive for Results. View our competency framework at: Here

UNICEF offers reasonable accommodation for consultants/individual contractors with disabilities. This may include, for example, accessible software, travel assistance for missions or personal attendants. We encourage you to disclose your disability during your application in case you need reasonable accommodation during the selection process and afterwards in your assignment.

UNICEF has a zero-tolerance policy on conduct that is incompatible with the aims and objectives of the United Nations and UNICEF, including sexual exploitation and abuse, sexual harassment, abuse of authority and discrimination. UNICEF also adheres to strict child safeguarding principles. All selected candidates will be expected to adhere to these standards and principles and will therefore undergo rigorous reference and background checks. Background checks will include the verification of academic credential(s) and employment history. Selected candidates may be required to provide additional information to conduct a background check.

Remarks

Individuals engaged under a consultancy will not be considered “staff members” under the Staff Regulations and Rules of the United Nations and UNICEF’s policies and procedures and will not be entitled to benefits provided therein (such as leave entitlements and medical insurance coverage). Their conditions of service will be governed by their contract and the General Conditions of Contracts for the Services of Consultants. Consultants are responsible for determining their tax liabilities and for the payment of any taxes and/or duties, in accordance with local or other applicable laws.

The selected candidate is solely responsible to ensure that the visa (applicable) and health insurance required to perform the duties of the contract are valid for the entire period of the contract. Selected candidates are subject to confirmation of fully-vaccinated status against SARS-CoV-2 (Covid-19) with a World Health Organization (WHO)-endorsed vaccine, which must be met prior to taking up the assignment. It does not apply to consultants who will work remotely and are not expected to work on or visit UNICEF premises, programme delivery locations or directly interact with communities UNICEF works with, nor to travel to perform functions for UNICEF for the duration of their consultancy contracts.

Company:

UNICEF

Qualifications:

Language requirements:

Specific requirements:

Educational level:

Level of experience (years):

Senior (5+ years of experience)

Tagged as: Classification, Data Visualization, Industry, NLP, Predictive Analytics, United States