LEAD Platform

• Problem at hand
• Solution
• What to build
• Design decision
• Build Path
  • Phase 0
  • Phase 1
  • Phase 2
  • Phase 3
• The Three Layers We Must Design Explicitly
  • Layer 1: Identity & Credentialing
  • Layer 2: The Initiative Schema
  • Layer 3: A Contribution Agreement

Asking the right question is a key. If we want to control the dev cycle and not let the cycle control us, we need to answer some questions before we start.

The platform is a system that empowers us to reach our goal. It will change in form through out different phases.

Problem at hand

Our country with significant economic potential and highly skilled, creative and authentic people. Held back by a weak economy, poor institutional support, and a leadership vacuum. Between 25% and 33% of the population now lives abroad.

Solution

Transform the economy from a consuming one to a productive one. Help local businesses find a place in foreign markets and start exporting using expats as the bridge.

What to build

A non-profit platform where local businesses post structured initiatives. Expats can contribute in defined ways: financial backing, legal guidance, logistics coordination, marketing, technical work, or direct mentorship

Design decision

• Initiative-lifecycle: it has states (draft → validated → active → fulfilled → archived)

• Reputation: the value a contributor/expat builds must be visible to future local businesses

Build Path

Phase 0

We should try to run actual process manually with 5-10 real local businesses and 10-20 real expats. Using WhatsApp, Google Forms, and a shared spreadsheet.

We will learn:

• How do local businesses actually describe what they need? To be able to create the correct schema
• What makes an expat commit?
• Where does the process break down when it’s manual?

Phase 1

Goal: Have real initiatives, real contributors, real outcomes with no custom code.

An openionated stack:

Phase 2

We can start introducing custom code only for the things that genuinely need it.

What actually needs custom code at this stage:

1. The Matching Engine — an expat in Dresden with logistics expertise in the EU food import chain should surface to a local food producer posting an initiative, not a lawyer in Toronto. Airtable filters can’t do this well. A Python service with a scoring algorithm can.

2. The Contribution State Machine — we need a defined workflow: expressed interest → matched → agreement signed → active → milestone verified → completed → reputation updated. Each state transition needs to be tracked. We can build a API endpoint (FastAPI for example).

3. The Reputation Ledger — every completed contribution must produce a verifiable artifact: a signed document, a confirmed milestone, a business rating. Expats who complete 3 initiatives should be visibly more trusted than those who completed 0. This is a database table, but it needs to be designed intentionally from the start.

What we should NOT build custom at this stage:

• Search
• Email/notifications
• File storage (use Cloudflare R2 free tier is massive)
• Video calls between expats and local businesses

Phase 3

Real usage data, a validated schema, a working matching algorithm, a contribution state machine, and a reputation ledger.

Stack recommendation for the custom build:

• Backend: Go

• Database: PostgreSQL been mirroring to it since Phase 1

• Search: Meilisearch, handles Arabic/French/local language characters well

• Auth: Keycloak

• Frontend: Next.js

The Three Layers We Must Design Explicitly

Layer 1: Identity & Credentialing

We need a tiered trust architecture:

Local Businesses:

• Tier 0: Self-registered (can browse, cannot post)
• Tier 1: Community-verified (another local business or a trusted local org vouches for them)
• Tier 2: Document-verified (business registration, tax ID, done asynchronously by a volunteer legal team)

Expats:

• Tier 0: Social media accounts verified (OAuth login, profile imported)
• Tier 1: Domain-verified (they’ve completed one initiative)
• Tier 2: Expert-designated (the community recognizes their specific expertise in a domain)

Layer 2: The Initiative Schema

This is the most important data design decision we’ll make. “Initiative” needs to be a rich, structured entity.

Layer 3: A Contribution Agreement

Every match between an expat and an initiative must generate a signed agreement in some form.

The agreement template should contain:

• What the expat commits to (from the contribution record)
• Timeline
• What constitutes “done” or a successfully finished contribution
• How disputes are handled
• Intellectual Property/ownership clause for any work product

Disclaimer LLM was harmed in the making of this article.