Built in Zambia, for learners everywhere.
Erva connects students with the right teacher through research-driven recommendations — and it's the first step toward a learning platform designed for the places that need one most.
The problems we're trying to solve
Three things shaped how Erva was designed.
Quality teaching is unevenly distributed
In Zambia and across much of the developing world, finding skilled teachers online is harder than it should be. Erva is built with those gaps in mind.
Search isn't matching
A long list of profiles doesn't help if you don't know which one fits. Erva surfaces teachers based on subject, style, language, and learning goals — not just keywords.
Teachers are people, not listings
Behind every profile is someone sharing what they know. Booking, messaging, and reviews are designed to keep the relationship at the center.
How Erva started
Erva began as a hobby — a small experiment in what a modern teacher-finding service could feel like. The name is a nod at Minerva, the Roman goddess of wisdom; trim the prefix and you're left with something shorter, friendlier, and easier to say.
Over time the project sharpened. As I dug into recommendation systems for my final-year computer science research, Erva became the testbed — a real platform where ideas about matching learners with the right teacher could be tried, measured, and improved. The hybrid model that powers recommendations today started as a homework problem.
What began as a side project is now a closed-beta marketplace: students can find teachers, book lessons, message in real time, and leave reviews. It's still early, deliberately small, and shaped by the people using it.
The platform, today
Real, working software in closed beta — not a roadmap.
Recommendation engine
A hybrid model — sentence embeddings plus subject, style, language, and rating signals — surfaces teachers based on what each student actually needs.
End-to-end booking
Students browse availability, request lessons, and join meetings; teachers manage calendars and confirm bookings. Timezone-aware throughout.
Real-time chat
Live messaging over WebSockets keeps students and teachers in contact before, during, and after a booking.
Honest reviews
Reviews are tied to completed lessons only. Ratings feed back into how teachers are ranked.
A learning platform for the places that need one most
Erva's longer-term goal is to grow from a teacher marketplace into a full learning management system — designed first for Zambia and other countries, where access to good online learning tools still lags far behind demand.
That means lighter pages, lower bandwidth, payment rails that work locally, and content that reflects the curricula and languages of the people using it. The marketplace is the seed; the LMS is what it's planted for.
About the founder
Hi, I'm Caleb. I'm the founder of this project, and the main developer of the pages you've seen so far. Erva was developed as the practical half of my thesis on recommendation systems for learning marketplaces, and is, for now, built single-handedly. I'm looking for a team to help bring the full LMS to the world.
Outside of Erva, I work on full-stack web projects, tinker with ML pipelines, and spend a lot of time thinking about how software solutions can serve people who've been an afterthought on most platforms.
Want to try it?
Erva is in closed beta. If you'd like to join — as a student or teacher — we'd love to hear from you.