Case study · LOEXA Africa
Building Africa's Workforce Capability Platform
LOEXA needed a digital platform to connect Africa's youth with skills training, industry mentors, and job opportunities — powered by AI matching. We built it from scratch in 4 months.
- Client
- LOEXA Africa
- Industry
- EdTech / Workforce Development
- Timeline
- 4 months
- Team
- 3 engineers + 1 designer
3,200+
Learners registered
47
Training orgs onboarded
93%
Screening time reduced
4
Countries launched
01 — Challenge
The problem worth solving
Africa has over 600 million young people under 25, but skills training programs are fragmented, hard to discover, and have no standardized credentialing. LOEXA wanted to unify this ecosystem — learners, training providers, and employers — on one platform with intelligent matching. The challenge: build something robust enough for enterprise but accessible enough for users on slow mobile connections across 10+ African countries.
02 — Solution
What we designed and built
We built a multi-sided platform: a learner-facing PWA optimized for low-bandwidth environments, a training provider dashboard to upload and manage courses, and an employer portal with AI-powered candidate shortlisting. The AI layer uses a fine-tuned matching model that ranks candidates by skills, location, and job relevance. All content is offline-cacheable for areas with unreliable internet.
AI-powered skills matching between learners and job opportunities
Offline-first PWA with service worker caching
Multi-role platform: learner, trainer, employer, admin
Course catalog management with video upload and progress tracking
Credential verification and digital certificates
Multilingual support (English, French, Swahili)
03 — Stack
Under the hood
- Frontend
- Next.js 14 (App Router) + TypeScript + Tailwind CSS
- Backend
- Django REST Framework + Celery for async job processing
- AI / ML
- Python (scikit-learn) + OpenAI Embeddings for matching
- Database
- PostgreSQL + Redis for caching and task queues
- DevOps
- AWS EC2 + S3 + CloudFront CDN + GitHub Actions CI/CD
04 — Outcome
What changed
The platform launched across 4 countries in the first phase. Within 6 weeks of launch, 3,200 learners had registered and 47 training organizations had onboarded their course catalogs. The AI matching system reduced employer screening time from 8 hours to under 30 minutes.
Tech Yugantar didn't just build what we asked — they challenged our assumptions, suggested better approaches, and delivered a product that genuinely works for our users. The AI matching alone is worth the entire project cost.
Want a similar result for your business?
Tell us your idea — we'll scope it, plan it, and build it.