SAIF initiatives connect talent, research, trusted data, infrastructure, and real problems through bounded work designed to produce credible and useful outcomes.
These focused themes support the broader mission; they do not define the full institutional portfolio.
SAIF's institutional framework for inclusive AI infrastructure development serving Syria and the Levant.
A phased framework for documenting, governing, and sharing representative data resources for responsible AI development.
SAIF's approach to transparent evaluation, reproducible benchmarking, and safety-aligned reporting for AI systems.
Initiatives are bounded pieces of research, capability-building, data, or institutional work. Status labels distinguish signed scope, open intake, development, and future work without presenting plans as completed delivery.
A long-running local knowledge and media archive covering communities across Syria. SAIF's signed agreement with the Syrian Computer Society identifies this material for study around responsible normalization, cleansing, cataloging, preservation, and possible future research use. This status does not imply an ongoing public deliverable.
Training of Trainers is identified as a potential joint activity under SAIF's signed collaboration agreement with the Syrian Computer Society. No cohort or delivery timetable is announced here.
A structured contributor pathway for data labeling, review, and preparation, supporting the quality of linguistic and knowledge resources under clear documentation, privacy, and review guidelines.
An intake pathway for people interested in research support, evaluation, documentation, trusted data, and other mission-aligned contribution. Submission does not imply an active placement or project assignment.
SAIF is developing research pathways across applied AI, trusted data, evaluation, and responsible adoption, with topics and titles to be announced only when confirmed.
Contributors, researchers, and institutions can review current pathways on Get involved or reach us through Contact.
