CASE STUDY 01 / AFFPILOT
Making scale feel effortless.
A monolith-to-microservices migration, faster background processing, and a more resilient platform for 85,000+ registered users.
01 / CONTEXT
The challenge
Affpilot’s AI content platform needed faster background processing and the ability to handle growing production workloads. The work spanned the application architecture, task execution, and cloud infrastructure.
02 / OWNERSHIP
My responsibility
As Tech Lead, I led the migration of Affpilot 2 from a monolithic application to microservices, redesigned the Celery task system, and diagnosed critical production issues.
03 / THE WORK
Engineering contributions
Separate services and workloads
Led the migration using Go, Python, RabbitMQ, gRPC, and Redis, moving the platform toward independently scalable services.
Improve the background execution path
Redesigned the Celery task system to reduce user wait times and improve parallel article generation.
Strengthen production delivery
Applied horizontal and vertical scaling on Azure VMs, deployed on Azure and GCP, and improved the content pipeline’s OpenAI prompts.
04 / THE BIGGER PICTURE
Architecture perspective
Service boundaries make independent scaling possible, while introducing communication and operational complexity. This is the architectural balance illustrated below; it is not a reconstruction of Affpilot’s internal system.
05 / IMPACT
What changed
- Reduced background-task wait times by 80%.
- Improved parallel article generation by 20%.
- Stabilized a platform with over 85,000 registered users and 3,000 active users.
Based on my professional experience as summarized in my resume. Figures are specific to the contributions described; underlying benchmarks and proprietary implementation details are not published here.
NEXT / ORBITAX