Harambee Youth Employment Accelerator Combats Youth Unemployment in South Africa with Google Cloud and AI

Harambee Youth Employment Accelerator, a South African not-for-profit social enterprise, is tackling the nation’s severe youth unemployment crisis, where over 50% of individuals aged 18–35 are jobless.

By harnessing Google Cloud and AI-driven solutions, Harambee has developed a scalable, data-driven platform that has connected over 4 million young people to job and skilling opportunities, dismantling systemic barriers to employment.

This case study explores how Harambee leverages Google Cloud technologies and AI to enhance job matching, reduce operational costs, and amplify its impact, transforming lives and offering a replicable model for addressing global youth unemployment.

Background

South Africa grapples with a profound youth unemployment challenge, with over 6 million young people aged 18–35 out of work and 44.7% of those aged 15–34 neither employed nor in education or training (NEET) as of Q4 2021.

Structural barriers such as limited formal work experience, inadequate access to education, high minimum wages, and exclusionary hiring practices exacerbate the issue. Employers often prioritize overqualified candidates to minimize risks, further marginalizing disadvantaged youth.

Founded in 2011, Harambee partners with government, private sector, civil society, and over 4 million youth to unlock earning opportunities, aiming to foster a growing economy powered by young people through innovative, data-driven solutions.

Challenge

Harambee faced significant hurdles in scaling its mission to address South Africa’s youth unemployment crisis. With over 9 million NEET youth and 1 million new labor market entrants annually, the organization needed a platform capable of reaching millions cost-effectively, particularly in underserved areas with low smartphone penetration.

Its on-premises infrastructure struggled to process and analyze data from over 1.4 million youth, hindering real-time insights. Traditional recruitment practices often excluded youth without formal qualifications or experience, necessitating a new approach to job matching that incorporated non-traditional skills and informal work experience.

Geographical barriers, such as high transportation costs and informal addressing systems, further complicated matching candidates to jobs within reasonable commuting distances. Additionally, systemic issues like exclusionary hiring practices and a mismatch between skilling programs and market needs limited opportunities for youth.

Solution: Leveraging Google Cloud and AI

In 2017, Harambee transitioned from on-premises data centers to Google Cloud, partnering with DotModus, a Google Cloud integration specialist, to build a scalable, cloud-native architecture. Supported by Google.org and Google for Nonprofits, this move aligned with Harambee’s mission and provided a robust technological foundation.

The SA Youth Platform (sayouth.mobi), a tech-enabled, multi-channel network, offers free access to job and skilling opportunities. This platform is hosted on Google Cloud’s App Engine and Kubernetes Engine for rapid deployment and scalability, while Cloud Storage and Cloud SQL manage large datasets, including candidate profiles and job opportunity data.

BigQuery and Looker Studio enable analysis and visualization of labor market dynamics, and Cloud Identity & Access Management ensures secure operations. Google Maps Platform facilitates geolocation services, matching candidates to jobs based on proximity to address transportation cost barriers. This infrastructure leverages Google’s Points of Presence in South Africa to reduce latency and costs.

Harambee employs AI and machine learning to enhance job matching by incorporating non-traditional metrics, such as informal work experience and behavioral attributes. Using TensorFlow and Cloud Dataflow, Harambee builds models to predict candidate suitability, factoring in skills like volunteering and metrics like job retention likelihood.

Apache Solr and PySpark index candidates and opportunities in BigQuery, creating a vector map model that predicts transport ease and matches candidates to jobs based on commute feasibility. Learning potential assessments identify 20% more suitable candidates overlooked by traditional metrics like school grades, countering exclusionary hiring practices.

The SA Youth Platform is mobile-first, accessible on any device, enabling candidates to apply without costly travel. Harambee’s partnerships with over 2,533 employer partners, government, and civil society, including initiatives like the Presidential Youth Employment Intervention, aggregate opportunities from 1,300+ partners.

Skilling programs like DigiLink, a 12-month work-integrated learning initiative, teach data analysis and software testing, transitioning participants to sustainable jobs. Mock interviews and the Youth Voices Programme further build confidence and engagement.

Results

Harambee’s adoption of Google Cloud and AI has delivered transformative outcomes. The organization has supported 4.1 million work-seekers, enabled 1.2 million opportunities, and placed over 50,000 youth in employment by 2024.

It expects to reduce data center costs by 70%, redirecting funds to strategic initiatives. AI-driven algorithms and geolocation have improved match accuracy, reducing commute times and costs, with over 400,000 youth attending work-seeker support sessions.

The learning potential assessment has opened opportunities for 20% more candidates, and partnerships with employers promote inclusive hiring. The cloud-based platform positions Harambee for global scalability, while its efforts have generated R30.7 billion in income for youth, contributing to economic inclusion. Harambee has identified 66,000 digital jobs, including 40,000 entry-level roles, and piloted programs to bridge skills gaps.

Key Takeaways

Harambee’s approach demonstrates the power of population-scale solutions, as noted by CIO Hanlie de Bod:

“When you want to solve population-scale problems, you need population-scale solutions, and Google Cloud has been able to provide that for us.”

Its dataset, likely South Africa’s largest on youth unemployment, drives evidence-based interventions. AI and alternative assessments challenge exclusionary hiring norms, unlocking opportunities for marginalized youth, while partnerships with Google, government, and private sectors amplify impact. Despite these achievements, Harambee faces challenges, including South Africa’s sluggish economic growth (0.1% in Q1 2025), skilling gaps, and the need for sustained funding to scale globally.

Moving forward, Harambee aims to enhance AI algorithms, expand digital economy initiatives, and deepen partnerships to create sustainable job pipelines and advocate for policy changes.

Conclusion

Harambee Youth Employment Accelerator showcases how technology, partnerships, and data-driven innovation can address complex social challenges. By leveraging Google Cloud and AI, Harambee has transformed South Africa’s youth employment landscape, connecting millions to opportunities and fostering inclusive growth.

Its model provides a blueprint for tackling youth unemployment globally, demonstrating the transformative potential of collaboration and innovation in building an inclusive society.