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, a connected world
UNICEF has a 70-year history of innovating for children. We believe that new approaches, partnerships and technologies that support realizing children’s rights are critical to improving their lives.
One of UNICEF's Information and Communication Technology Division (ICTD) strategic goals is to transform and build partnerships to successfully implement UNICEF programmes globally through the use of innovative technology-enabled solutions.
UNICEF’s ICTD and Office of Innovation are working together to implement R&D capacity on Big Data, AI and Data Science at UNICEF's core through the MagicBox project. This project is transitioning from 4 years of early stage development in the Office of Innovation to the recently-established Field Solutions Unit at ICTD: a team of experienced software developers, data scientists, and designers within ICTD’s Solutions Centre and Support, working hand by hand with the Office of Innovation on a daily basis.
The Unit’s goals are to establish enterprise support to innovative open-source tools used by field-based programmatic experts, planning specialists, partnership managers, and implementing partners around the world. They work with counterparts throughout the organization to tailor open-source software solutions that magnify UNICEF’s impact by driving better decision-making, promoting collaboration, and realizing organizational efficiencies with modern digital tools.
How can you make a difference?
We're an interdisciplinary global team tasked with identifying, prototyping, and scaling new technologies and practices. With our partners, and through R&D cycles, we focus on convening and collaborating on new and different solutions using Big Data, AI and Data Science to:
- Accelerate adoption of data-driven decision-making practices by programme officers to achieve results for children more quickly and more efficiently;
- Provide operational insights to programme managers, section chiefs, and country management teams to allow managers to better prioritize their attention/efforts, identify problematic partners, highlight data gaps, improve staff accountability, and to respond more effectively to changing contexts;
- Enable adaptive management of programmatic interventions, projects, pilots, and innovations for continuous improvement, course-corrections, risk mitigation, and feedback loops;
- Enable data exploration and discovery, anomaly detection, identification of correlations across disparate data sets, and discovery of new tracer indicators to develop predictive analytic capabilities within the organization;
- Realize greater value from existing data sets by contextualizing with other public, institutional, or otherwise-accessible data sets related to the data’s programmatic context.
We are looking for a Machine Learning Lead to join the Science team of MagicBox - a team that does research and support product development with non-traditional data (i.e., Big Data) and methods (i.e., computational methods). It does so by combining several disciplines such as network science, complex systems, natural language processing, data mining, machine learning, remote sensing and artificial intelligence looking into programmatic problems such as epidemiology, sudden onsets, urbanization, migration, human dynamics, vulnerabilities, poverty, inequalities and bias.
The Magic Box Team collaborates with world class scientists, on the ground responders, private sector data holders and tech leaders to push the Data Science and AI agenda towards the problems that affect the most vulnerable children. Breakthroughs and working methodologies are streamlined to the Development Team to scale into operational tools that can make a real impact.
You will contribute to the development of Machine Learning and Data Science open source projects and research that support country offices and UNICEF Divisions, while staying current with the latest technologies, applying them according to best practices.
The exact tasks for the consultant will be jointly determined in an agile way and are outlined in summary below. The consultant will directly interface with the Chief Scientist and Product Manager of Magic Box and will work on tasks/areas as prioritized and authorized by the Chief Scientist each month. Upon completion of tasks, the Chief Scientist will certify completeness and quality of work for payment.
Ongoing initiatives that the candidate will be expected to support:
- Vulnerabilities and socio-economic indicators: Combining different computational techniques and data sets to estimate socio-economic indicators such as poverty, access to information or resilience to exogenous events (e.g., natural disasters).
- Inequalities and bias: Develop techniques and methodologies to identify, quantify and correct bias and inequalities in data and algorithms to ensure that the most marginalized are represented.
- R&D: Exploration of new venues by utilizing novel computational techniques and applications of Machine Learning, Data Science and Network Analysis and through collaborations with academic and private partners.
Main Responsibilities will be:
- Point of contact for ongoing initiatives between the MagicBox team and the main actors in the field of machine learning (academia, private sector, other UN agencies).
- Develop and manage relationships with partners relevant to the ongoing initiatives: identify partners that are developing methods and/or cutting-edge research in the key area of support. Find opportunities to collaborate/co-create/integrate machine learning efforts in UNICEF programmatic areas. Identify stakeholders within UNICEF and other UN organizations to build collaborations for applying research to UNICEF goals. Work on producing joint research and analysis resulting in peer reviewed publications, reports, and working methodologies.
- Produce high impact research: working with colleagues in UNICEF, partners in academia and private sector to produce high impact scientific work. Conduct scientific analysis and strengthen the research collaboration network.
- Advice, mentor and oversee visiting researchers and junior data scientists.
- Build quick prototypes: Produce quick analysis/prototypes that allow a technical exploration of a data science concept to country offices and partners.
- Support integration of prototypes into operations: support as technical expert the integration of the validated prototypes in the country office operations, putting together technical partners and programme specialists.
- Provide technical assistance to the Innovation Fund, Country Offices and start-ups, including reviewing related applications to the Innovation Fund, specifically in the area of machine learning.
- Documentation and knowledge management: Maintain overall portfolio documentation, including lessons learned of the managed projects on the key area of support.
Required skills and experience:
- PhD degree and at least 4 years of professional/research experience in a relevant field (Data Science, Computer Science, Applied Mathematics, Physics, or another comparable field).
- Experience in communicating technical concepts to diverse audiences
- Attention to detail, research skills and the ability to work under strict deadlines is a must.
Experience and Knowledge required for the assignment:
Experience with the following:
- Strong Machine learning and Data Science knowledge (deep understanding of methods and their appropriateness of modeling, supervised and unsupervised techniques depending on data conditions and contexts). Knowledge on CNN/deep learning is a plus. Network Analysis and Complex Systems as main areas of research, specially applied to Computational Sociology. Strong record of scientific publications showing the above and working experience.
- Past experience in working with behavioral data including mobile phone traces, twitter data, email interactions, or similar datasets is preferred. Experience with working with very large datasets is a plus.
- Visualization techniques for producing eye catching images and plots (including shapefile/raster mapping)
- Good research record (scientific publications related to the professional and academic experience required). Research experience focused on algorithmic and data bias and working with sensitive datasets is a strong plus. General approach to research and development with attention to the privacy and ethics of data and algorithms.
Software/development environments experience:
- Experience with cloud platforms, familiarity with Microsoft Azure and Spark - as our corporate platform for scientific computing- is an asset and knowledge of others (AWS, Google Cloud) a plus.
- Programming languages: Python and/or R a must.
- Open values and general understanding of open source (code, data, platforms, initiatives, etc.) is a must. Specific experience/knowledge on working with/on open source initiatives a plus.
- Prior experience working with a multilateral or UN organization is highly desirable
Language Requirements: Fluency in English is required. Additional UN languages will be considered an asset.