Context
Operational context
Multiple apartment complexes were managed through spreadsheets, telephone calls and paper records. Financial, maintenance, tenant and asset information lived in separate workflows, making it difficult to see the full operational picture.
Challenge
The problem
Disconnected processes caused delayed rent collection, inconsistent reporting, tax calculation errors and limited visibility into maintenance and asset condition. Repair histories and spatial utility records were difficult to trace, contributing to service delays, unplanned costs and poor tenant experiences.
Constraints
- Financial, tenant and asset records needed to be unified without losing operational accountability.
- Payment and tax workflows required accurate, secure processing.
- The platform needed to work well for both management and tenant-facing use on desktop and mobile.
Architecture
System design
Django and PostgreSQL provide the core application, business rules and REST APIs. A responsive Bootstrap and JavaScript interface supports management and tenant workflows. Scheduled jobs drive billing, reminders and maintenance alerts, while Pesapal handles online payments. The platform runs on AWS in Docker and Kubernetes behind Traefik and cert-manager, with Jenkins, Git and Ansible supporting CI/CD and Prometheus, Grafana and Alertmanager providing operational monitoring.
Solution
What was built
I designed and built a central platform for rent, billing, taxes, maintenance, stores, tenant communication, spatial as-built plans and property assets. Automated reminders, online payment integration, task assignment and operational dashboards replaced disconnected manual steps and gave managers a reliable system of record.
Engineering judgement
Decisions and trade-offs
Unified operational model
Context: Financial and maintenance records were disconnected
Decision: Use one PostgreSQL-backed domain model with role-based modules
Trade-off: A broader model required careful boundaries between teams
Scheduled workflow automation
Context: Billing and reminders were repetitive and time-sensitive
Decision: Run controlled background schedules for billing cycles, reminders and alerts
Trade-off: Schedules require monitoring and idempotent jobs
Containerised cloud deployment
Context: The platform needed reliable updates and secure routing
Decision: Deploy containers on Kubernetes with Traefik, certificates and automated CI/CD
Trade-off: The infrastructure adds complexity but makes releases repeatable
Contribution
My role
- Mapped property management workflows and designed the system architecture.
- Built the Django backend, PostgreSQL data model and REST APIs.
- Developed responsive management and tenant interfaces.
- Integrated billing schedules, maintenance alerts and Pesapal payments.
- Deployed and monitored the platform on AWS with Docker, Kubernetes and CI/CD automation.
Product demonstration
See the system in use
Results
Operational impact
- Improved maintenance response through real-time issue reporting and task assignment.
- Made asset repairs, depreciation and lifecycle information easier to track.
- Improved billing transparency, tenant communication and service delivery.
- Provided management with current financial and operational information.
Reflection
Lessons learned
- Operational platforms are strongest when finance, maintenance and asset workflows share identifiers and ownership rules.
- Automated financial workflows need visible exceptions and reconciliation paths.
- Monitoring scheduled jobs is as important as monitoring the web application.