Case study
Northline: running a UK last-mile delivery operation on one platform
From the first contact with a driver candidate to the weekly pay run: recruitment, compliance, scheduling, fleet, finance and a driver app, built and modernised by Malvixia.
01 — The client
Northline is a UK last-mile logistics operator. It runs delivery routes for a major e-commerce delivery network from multiple depots, with a pool of self-employed delivery drivers, vans and regional managers.
02 — The problem
The whole driver lifecycle had to work every single week: finding and onboarding drivers, keeping licences and documents valid, putting the right driver in the right van at the right depot, tracking damage and fines, and paying everyone correctly against weekly performance scorecards.
Spread across spreadsheets, inboxes and separate tools, that means retyped data, missed expiries, slow onboarding and pay disputes. Northline needed one system that operations, HR, compliance, fleet and finance teams all use daily, plus an app for the drivers themselves.
03 — What we built
An internal operations platform with role-based access for 15 user types, from recruiters and trainers to depot, area and regional managers, fleet and garage staff, finance and drivers.
- Driver recruitment & onboarding: lead capture from job boards, social channels and driver referrals; candidate pipeline, invitations, training and onboarding steps.
- Compliance: document collection and expiry tracking, UK driver-licence checks and an audit trail on sensitive records.
- Depots, regions & agencies: depot and region structure, depot groups, pay rates, job titles and agency records.
- Scheduling (DA rota): daily and weekly scheduling of delivery associates, calendars, availability, and check-in / check-out.
- Fleet & garage: vans, vehicle inspections with check-out and check-in sheets, damage records, garages and GPS tracking integration.
- Finance: invoices, weekly self-billing, deductions, deposits, driver charges, end-of-contract balances and PDF documents for every step.
- Performance incentives: import of weekly delivery-partner scorecards and automatic calculation of driver scores and incentive payments.
- Fines (PCN) management: penalty notices matched to drivers and vehicles.
- AI document extraction: supplier invoices are read by an AI model, which extracts line items, amounts and vehicle registrations, with amount verifiers and an AI log so finance can review each result before it is used.
- Internal messenger: chat between office teams and drivers, plus notifications and task management.
- Driver Hub & reports: manager-driver meetings, operational reports and dashboards.
- iOS & Android driver app: a cross-platform React Native app for drivers, plus a mobile web app.
04 — Modernisation: the new platform
Alongside the live system, we designed and started a new, hardened platform so the business can grow without a risky big-bang rewrite:
- A full technical audit of the existing system: architecture, data model, integrations, security, testing and migration readiness.
- A strangler-fig migration plan, module by module, from MySQL to PostgreSQL.
- New backend services in Java 21 / Spring Boot with tenant isolation, signed-token authorization, an outbox/inbox messaging pattern on RabbitMQ and managed schema migrations.
- A new React / Vite frontend with permission-guarded, lazily loaded modules and end-to-end tests.
- Containerised infrastructure with Docker Compose, an Nginx gateway, health checks, CI pipelines and clean-room deployment tests.
05 — Tech stack
- Live platform: Python, Django 4.2, Django REST Framework, MySQL, Celery, Redis, S3-compatible object storage, Next.js, React, MUI.
- Driver app: React Native (Expo), TypeScript; mobile web app on Next.js.
- New platform: Java 21, Spring Boot 3, PostgreSQL, RabbitMQ, Redis, React, Vite, TypeScript, Playwright, Docker, Nginx.
- AI: LLM-based document extraction with validation and human review.
06 — Scale of the work
- 1,000+ commits across six repositories between October 2025 and September 2026.
- 19 backend modules, around 217 data models, around 388 API routes, 48 background tasks and 27 scheduled jobs.
- 15 user roles, with change history kept on 83 data models.
- Live system in daily use, with continuous releases while the new platform was being built.
07 — Timeline
- October 2025: first commits of our engagement on the live platform.
- November 2025 – March 2026: driver app and mobile web app; steady feature work across recruitment, scheduling and finance.
- April – September 2026: peak delivery, with 150–200 commits a month: finance, fines, fleet sheets, messenger, AI invoice extraction.
- August 2026: technical audit, migration plan and the new platform foundation (backend, frontend and infrastructure).
08 — Ready to talk?
Running operations on spreadsheets and disconnected tools? Name the problem. You get a yes, a no, or a date, not a sales sequence.
Book a 30-min call · or start with the free AI & Automation Readiness Audit.