AdTech · Fintech · R&D
Trading Desk & Billing
from monolith to scalable architecture
$40K+
Daily Revenue Processed
5M+
Daily Aggregated Records
100+
DSP Integrations
10-min
Aggregation Pipeline
Context
The platform is a billing system for DSP partners.
Problem
Initially the system was built as a single Laravel monolith.
- Analytical queries competed with transactional operations for resources.
- A shared runtime reduced resilience: API degradation could take down the entire billing system.
- Scalability was limited without putting financial calculations at risk.
Project goals
- increase the speed of statistics collection and processing;
- extend discrepancy calculation with new analytical dimensions (day, hour, country, spot);
- separate analytical and transactional workloads;
- isolate external APIs from internal platform services;
- improve security, resilience, and system scalability;
- preserve financial calculation correctness and compatibility with existing billing.
My role
During implementation, the project team was fully under my management.
Analysis & Planning
Identifying architecture bottlenecks and forming functional and non-functional requirements.
Delivery Management
Coordinating the cross-functional team, prioritizing the migration backlog, controlling TTM.
Project results
The project was successfully launched to production.
For the business
- discrepancy could be calculated across multiple analytical dimensions;
- reporting was built faster;
- financial calculation correctness was preserved;
- a scalable foundation emerged for further analytics services development.
For the platform
- analytical and transactional workloads were separated;
- clear areas of responsibility emerged between services;
- architectural review and change management processes were introduced;
- the platform became easier to operate and evolve.