Land-plot marketplace with a next-step plan: MVP in 2 months

The task

The client was launching a marketplace for land plots and the services needed to develop them. A buyer who finds a plot does not know which works are mandatory — title check, surveying, geology, utilities — and hunts for contractors by hand, while contractors get leads with no context. All that existed was a visual prototype: a map with filters, no backend, roles or requests.

What we built

• Built the full platform: backend, map with clustering and radius search, plot catalogue with filters and address geocoding. Runs in containers over HTTPS; the live stand was updated with no data loss.
• A rule-based next-step engine at the core: from plot attributes, deal type and land category it returns a prioritised list of works, each with a stated reason. No ML; rules change without retraining.
• Recommendations turn into a development plan with step statuses in one action. Contractors are matched by service and region; a request captures contacts and consent and notifies the company.
• Three dashboards — user, company, admin — with company verification and listing moderation. Personal-data compliance: consents, legal documents run by admins, seller contacts hidden from guests.
• Laid the groundwork for importing plots from external sources (base importer, three adapter types, job log with deduplication); real sources are connected in the next phase.

Outcome

2 months

from approved spec to pre-release MVP on an HTTPS stand — week 9 against the plan's own 10–13 week estimate

7 days

from spec to the first full-implementation commit: 168 files, 21.5k lines, then two months of hardening to pre-release

93 endpoints · 26 tables · 19 screens

MVP scope: 13 domain modules, 82 tests, 5 roles (three dashboard types plus moderator and super-admin)

11 rules

next-step engine: prioritised list of works, ~20 service types across 14+ categories, a reason for every recommendation

More cases