brain
projects·project·shipped

Bloom

Thesis

Multi-tenant apprenticeship platform for African markets. Replaces the WhatsApp-groups-plus-manual-bank-transfers status quo with structured cohort management, real-time collaboration, and automated split payments. Co-founded; Zaid led engineering.

Status

shipped (portfolio: Live). 1,200+ users, 50+ live programs, 15+ partner organizations onboarded within three months.

Stack

  • DB / auth / realtime / storage: Supabase (PostgreSQL).
  • Multi-tenancy: Row-Level Security policies — data isolation enforced at the database layer, no application-level filtering.
  • Messaging: Supabase Realtime + PostgreSQL LISTEN/NOTIFY for low-latency in-cohort communication. Messages persisted and searchable.
  • Payments: Paystack Split Payments API — automatic apprentice→mentor payouts with platform-fee extraction, compliant with regional regulations.
  • Permissions: PostgreSQL triggers enforce role-based access (mentor / apprentice / admin) at the database, not in application code.

Key decisions

  • RLS over application-level multi-tenancy: moves access control into the DB, eliminating a whole class of isolation bugs. Cost: complex SQL policies. Worth it.
  • Supabase over custom backend: auth + storage + realtime + Postgres in one stack. Lock-in risk accepted for shipping speed.
  • Split payments over manual invoicing: removes the single biggest operational bottleneck for partner organizations.
  • Permission model in the database, not the app: same principle as RLS — trust the strongest enforcement layer.

Learnings

  • Onboarding is the product for markets with varying technical literacy. Twelve iterations to reach 70% completion rate. Every step had to work without a tutorial.
  • Payment compliance is country-specific across African markets — different fee structures, reporting, regulations. Abstract early or pay later.
  • Moving access control into the database is one of those "pay the upfront complexity cost to avoid a class of bugs" decisions that compounds.

Outcomes

  • 1,200+ users across 50+ live apprenticeship programs.
  • 15+ partner orgs onboarded in first 3 months.
  • Validated the thesis that structured apprenticeship delivery can work at scale in markets where informal training is the norm.

Open questions

  • Where is the product today? Portfolio lists Live but no repo surfaced in this corpus (public source absent).
  • Unit economics: is the platform fee sustainable across markets with different payment-processing costs?
  • Retention: after the first cohort, do partner orgs stay?

Links