Job Opportunity: Data Scientist - Computer Vision

News

Job Opportunity: Data Scientist – Computer Vision

March 23, 2026

Please kindly read the entire job posting. Application instructions are located at the end.

Data-Pop Alliance is looking for a Data Scientist for the implementation of the “Creation of New and Multi-Modal, Multi-Lingual, and/or Multi-Cultural Datasets for More Equitable AI,” project in partnership with Mozilla. We are an international nonprofit created in 2013 out of the Harvard Humanitarian Initiative (HHI), the MIT Media Lab and ODI. We bring together researchers, experts, practitioners, and activists to change the world with data through three pillars of work: diagnosing local realities and human problems with data and AI; mobilizing capacities, communities, and ideas towards more data-literate societies; and transforming the systems and processes that underpin our societies and countries.

DPA has partnered with and received funding from organizations such as the Inter-American Development Bank, UNDP, FAO, UN ESCWA, GIZ (German development agency), the Vodafone Institute, Oxfam México and other key international stakeholders to develop projects in Sub-Saharan Africa (SSA), Latin America and the Caribbean (LAC), the Middle-East and North Africa (MENA), and Asia. To learn more about DPA, please consult our website and our “Overview and Outlook 2022-2024” report.

DPA’s projects are managed and carried out by a core team of about 30 people, including Directors, Managers, Officers, Data Scientists, Research Assistants, and Interns.

Dates: Immediate start. This is a fixed-term position linked to the implementation of the Mozilla project, expected to run from March 15, 2026 to May 31, 2026, subject to project funding and performance.

Location: Remote.

Scope of Work

The Data Scientist will report to the Data, Technology and Innovation Director and the Technical Manager and collaborate with the tech team. The role will support the implementation of activities and deliverables associated with the Mozilla project, in accordance with the project timeline and objectives.

Computer Vision and Data Processing Pipelines

1. - Support the design and technical structuring of the multimodal dataset, ensuring proper integration of video recordings, time-aligned text annotations, and metadata.
  - Define data formats, storage structures, and documentation standards to ensure interoperability and usability for AI research and development.
  - Computer Vision and Data Processing Pipelines
  - Develop and implement preprocessing pipelines for video data, including frame extraction, normalization, and quality control procedures.
  - Apply computer vision techniques to support sign language analysis, including feature extraction, pose estimation, or motion tracking where relevant.
  - Ensure that the dataset structure supports future applications such as sign language recognition, sign-to-text translation, and multimodal alignment.

Multimodal Alignment and Annotation Support

1. - Support the integration of video data with time-aligned textual annotations (e.g., SRT files) including gloss-level annotations and Portuguese translations.
  - Assist in validating annotation consistency and temporal alignment between modalities.

Data Quality Assurance

1. - Develop and implement procedures for dataset validation, quality control, and consistency checks across modalities.

Dataset Delivery

- Support the final preparation, packaging, and documentation of the dataset to ensure it meets the technical requirements for delivery to the Mozilla Data Collective.

Profile/Qualifications

Master’s or PhD in Computer Science, Data Science, Artificial Intelligence, Machine Learning, Computer Vision, Statistics, Applied Mathematics, or a related quantitative field.
Additional certifications or coursework in deep learning, computer vision, or AI are considered an asset.
Experience developing computer vision models, including tasks such as:
- Image classification
- Object detection
- Image segmentation
- Feature extraction
Experience with deep learning architectures used in computer vision (e.g., CNNs, ResNet, YOLO, Vision Transformers).
Ability to process and manage large image or video datasets.
Ability to work independently and as part of a diverse, multicultural team.
Bilingual proficiency in English is requisite (able to write and deliver conferences, reports, etc in both languages). Portuguese or Spanish are highly desirable.
A self-starter, disciplined, driven, eager to learn, grow, and make an impact.

Remuneration

The selected candidate will receive a monthly remuneration of 2,500 USD, for a total amount of 6,250
USD, payable in accordance with the organization’s payment procedures.

Recruitment Process

Qualified candidates are requested to submit 1) a CV, 2) a Cover letter including contact details for two academic or professional references to jobs@datapopalliance.org with the subject line “R2026001_DS” by March 30, 6:00 pm EST. Candidates will be reviewed on a rolling basis, so early application is strongly encouraged.

Please note that only complete applications submitted in English will be considered. Due to the large volume of applications, only short-listed candidates will be contacted. Also, rename each document you’ll submit with your name, the position and the type of document you are submitting. Ie. FIRSTNAME LASTNAME_Position_CV and send them in PDF format.

Equal Opportunity

Data-Pop Alliance employs personnel without regard to race, ancestry, place of origin, color, ethnic origin, language, citizenship, creed, religion, gender, sexual orientation, age, marital status, physical and/or mental handicap or financial ability. While remaining alert and sensitive to the issue of fair and equitable treatment for all, Data-Pop Alliance has a special concern with the participation and advancement of members of four designated groups that have traditionally been disadvantaged in employment: women, visible minorities, aboriginal peoples and persons with disabilities.