Scaling Azerion’s AdTech Platforms Through Strategic Staff Augmentation
Client
Azerion
Industry
AdTech
Location
Netherlands
Projects Delivered
20+
Team Size
45+
Service Used
Engagement Length
2019 - Ongoing
About Client
Azerion is a Dutch technology company founded in 2013. They operate one of Europe’s largest platforms for digital advertising and online games.
The company connects advertisers with audiences through mobile apps, websites, and social games. They focus on making the process of buying and selling digital ad space simple, safe, and efficient.
Project Overview
Azerion teamed up with Vivasoft in 2019 to boost their development speed and expand their engineering reach. What began as a small collaboration started with Vivasoft providing one developer for one specific project. This first step allowed both companies to see how well they could work together.
The results were positive. Vivasoft’s team blended into the daily workflows and met the technical demands of the AdTech world. As Azerion took on new challenges, they asked Vivasoft for more support. The team grew steadily from one engineer to a group of over 45 engineers.
Over the years, this partnership covered more than 20 different projects. The engineers worked on everything from media tools to advertising platforms. Today, Vivasoft remains a key part of Azerion’s development strategy. What began as a simple request for extra hands has turned into a deep, years-long engineering bond.
“Since partnering with Vivasoft, our team has been able to automate many of the up our time to focus on more important tasks, such as developing new products and services.”
Evelyn Hayes
CEO at azerion
Team Augmentation Journey
Total Resources
45
Frontend Developers
Project Managers
SQA Engineers
UI/UX Designers
Backend Developers
1
1st Concept
JAN 2019 – Ongoing
With our 1st project with Azerion, we have provided dedicated engineering support for the core advertisement server. Helped screen and manage global ad placements for B2B and B2C businesses in real time. And ensured the platform remains stable and accurate across a vast network of digital sites.
01
Software Engineer
2
2nd Concept
Nov 2021 - Ongoing
Vivasoft rebuilt the legacy AdTech infrastructure to resolve years of technical debt. We have modernized the backend stack, simplified massive database structures, and implemented a fresh user interface. This transition ensures the platform remains stable, secure, and easy to maintain for future business needs.
01
Team Leader
01
Web Application Architect
03
Backend Developers
03
Frontend Developers
Multiple
Project Managers (Client Side)
01
SQA Engineer
3
3rd Concept
Jul 2023 - Dec 2025
We have managed high-performance real-time bidding and unified ad operations through a Go-based monorepo. Connected multiple external platforms into a single system with over 20 microservices. Ensured sub-50ms response times and total data accuracy for global ad auctions.
01
Senior Go Developer
01
Solution Architect
4
4th Concept
March 2025 - Ongoing
Vivasoft joined this project in early 2025 to build a central hub for artificial intelligence. The work began by creating a unified API and a secure billing system to handle thousands of concurrent users. Since then, the team has expanded the platform to include an automated marketplace for AI tools. Currently, Vivasoft is working on more advanced features to help large companies manage their data and AI workflows securely.
02
Senior Backend Developers
01
Frontend Developer
01
DevOps Engineer
01
Python Engineer
01
Project Manager
Our Industry Expertises
Our deep understanding of diverse industries empowers us to design customized software solutions. Our engineers used the following technologies while working in the projects:
Javascript
Spring
python
java
Laravel
angular
react
MySQL
Amazon AWS
Bamboo
Grafana
Cloudwatch
php
Docker
Prometheus
Redis
Kotlin
C++
Lua
Next.js
AWS ECR
Typescript
AWS ECS
Golang
A multi-year partnership delivering 20+ digital projects
Projects
01/20
- Engineering Scalability
- Platform Modernization
- High-Performance AdTech Engine
- Unified AI Hub
01/04
Ad Serving Platform: Engineering Scalability
Summary
Iceberg is the foundational ad server for Azerion’s digital advertising strategy. Vivasoft provided specialized engineering expertise to help Azerion’s in-house team manage and screen global ad placements. This partnership began in 2019 and continues today. Vivasoft ensures the server remains stable and efficient as it processes massive volumes of B2B and B2C advertisements.
Duration
5+ Years
Industry
AdTech
What was done
Core Ad-platform Enhancement, Feature Implementation, and Production Engineering Support
Challenges faced by the Client
Azerion needed to strengthen its internal development team to maintain its core advertisement server. The platform requires a very high level of accuracy to screen and place ads correctly across a vast network.
Finding engineers with specific AdTech knowledge was a challenge. They needed an expert who could integrate quickly and assist with complex implementation tasks without disrupting daily operations.
Solutions Provided by Vivasoft
Vivasoft provided a dedicated software engineer with deep experience in AdTech solutions to act as an extension of Azerion’s team. This support allowed the client to accelerate their development cycles and improve the server’s screening process.
- Seamless Team Integration
The Vivasoft engineer joined the internal team to help manage the workload. This extra capacity ensured that new features and updates were delivered on time. It also allowed Azerion’s senior staff to focus on high-level strategy while Vivasoft handled critical implementation details.
- AdTech Technical Support
Vivasoft used its industry-specific knowledge to assist with the meticulous screening process of the Iceberg server. By applying best practices in ad delivery, the engineer helped ensure that placements were effective for both B2B and B2C clients. This collaboration kept the server running smoothly during periods of high traffic.
- Long-Term Continuity
Since 2019, Vivasoft has provided consistent support for this project. This long-term commitment has created a deep understanding of the Iceberg codebase. It ensures that the ad server remains a reliable cornerstone of Azerion’s advertising ecosystem as technology evolves.
02/04
Platform Modernization
Summary
Azerion’s demand-side platform and admin platform, known as Origin, had a decade-old burden of technical debt. The systems relied on a 10-year-old Spring Platform alongside a Java version that lagged behind three major releases. It turned the maintenance of the system difficult, prone to errors as well as resource-heavy.
It was quite clear that both the backend infrastructure and the frontend experience needed modernization. With a rigorous audit, a tailored solution was implemented to fix all the issues disrupting their system to transform the platform.
The final result was the stable, modern platform that is easy to maintain with a modern user interface that’s ready for years of growth.
4+ Years
AdTech
Full-Stack Legacy Migration & UI Overhaul
Challenges faced in this Project
- Accumulated Technical Debt Across Core Platforms
The Demand Side Platform (DSP) and Admin platform (Origin) had accumulated years of unresolved technical debt, making maintenance, feature development, and stability increasingly difficult.
- Severe Database Redundancy
- Faulty Event Management Architecture
- Outdated Backend Technology Stack
- Minimal Automated Test Coverage
The DSP and SSP platforms had only 200–300 automated tests, making deployments risky and increasing the likelihood of production issues.
- Disorganized and Obsolete API Structure
Hundreds of outdated APIs cluttered the codebase, while many active APIs were implemented in incorrect microservices, violating architectural boundaries, and increasing confusion.
- Manual and Inefficient Deployment Process
- Lack of Centralized Technical Documentation
Solutions
- Comprehensive Technical Debt Reduction
Conducted a systematic review of the DSP and Admin platforms to identify, prioritize, and resolve critical technical debt areas impacting stability and development efficiency.
- Database Cleanup and Documentation
Removed over 200 unused database tables and thoroughly investigated the remaining 600+ tables, documenting their purpose and identifying areas for optimization.
- Event System Redesign
Rebuilt the event management architecture to eliminate circular dependencies and prevent uncontrolled event growth, restoring system stability.
- Modernization of Backend Stack
Upgraded all microservices from the outdated Spring Framework to the latest version and migrated Java to the current supported release, improving security, maintainability, and development speed.
- Automated Testing Expansion
Expanded automated test coverage from approximately 200–300 tests to over 1,500 tests, significantly improving code quality and deployment confidence.
- API Cleanup and Reorganization
Removed hundreds of obsolete APIs and reorganized active endpoints into their appropriate microservices, restoring architectural clarity and simplifying maintenance.
- Containerization and Orchestration
Implemented Docker and Kubernetes to streamline deployments, replacing the previous Eureka-based service discovery with Kubernetes-native service management.
- Comprehensive Technical Documentation
Created detailed wiki documentation covering system architecture, historical issues, implemented solutions, and database structure to preserve technical knowledge and support ongoing development.
UI Screens
Challenging Features of the Software
Angular 13 - Compatible Authentication Module
Upgraded the shared pol-auth authentication library from Angular 9 to Angular 13 through a separate Git branch, enabling secure authentication support for the Admin UI while preserving backward compatibility for existing platforms.
Custom Geo Targeting Selector (Region & Country Level)
Developed a fully custom, multi-layer geo-targeting selector from scratch, allowing users to select entire regions or individual countries with precise add/remove controls on both available and selected lists.
Custom Size Targeting Selector
Built a custom size-targeting selector without third-party libraries, enabling users to manage size selections dynamically with dual-list add/remove functionality.
Persistent Search & Filter State
Implemented persistent search and filter functionality for selected modules by storing filter states in local storage, ensuring user selections remain intact across page navigation and refreshes.
Auto-Updating Job List with Manual Refresh Option
Implemented an auto-refresh mechanism that updates bulk upload job lists every 15 seconds, along with an instant manual refresh option to ensure real-time visibility of job status.
Dynamic List & Form Management on a Single Page
Designed a non-standard CRUD interface where lists and forms coexist on the same page, supporting create and edit actions with real-time list updates. Implemented advanced Redux state handling to prevent stale data and ensure consistency.
Dynamic Data Center Configuration with Hover-Based Controls
Created a dynamic data center list featuring hover-activated dropdowns that reveal configurable fields based on user interaction, reducing visual clutter and improving usability.
Rate Card Validation with Intelligent Redirect Logic
Enforced business rules by validating publisher–placement combinations during rate card creation. Automatically redirected users to the existing rate card edit page when duplicates were detected, preventing invalid configurations.
03/04
High-Performance AdTech Engine
Summary
Vivasoft built Ad Exchange, a high-performance AdTech engine to unify fragmented ad operations for Azerion. The project uses a Go-based monorepo to handle real-time bidding and content delivery. Vivasoft solved critical latency and data sync issues using a custom microservices architecture. Currently, the system manages over 20 services, providing a fast and resilient environment for global digital advertising.
2+ Years
AdTech
Cloud-Native Microservices Development
Client’s Business Challenges
Azerion struggled to manage growth as its AdTech ecosystem became more complex. They needed a way to handle high-speed bidding without losing data accuracy across different platforms. Such as:
- Disconnected Platforms
The team had to manage campaigns across multiple DSPs. These systems did not talk to each other. This fragmentation made it difficult to keep advertising data consistent across the entire network.
- Extreme Speed Requirements
In the real-time bidding market, every millisecond counts. Any delay in the system caused lost revenue and unfilled ad slots. Azerion needed a solution that could respond to bid requests almost instantly to stay competitive.
- Complex Task Management
Running multiple servers at once created technical conflicts. Sometimes, different instances tried to perform the same task simultaneously. This led to race conditions and made it hard to keep the background jobs stable.
Solutions
Vivasoft designed and engineered a high-performance monorepo system called Ad Exchange. They built this solution using the Go programming language to handle heavy traffic and complex ad operations. The system unifies several external platforms into one reliable and fast environment.
- Reliable Data Synchronization
Vivasoft used a custom Saga Pattern to keep data accurate across different platforms. This system acts as a state machine for every transaction. If an update fails on an external site, the system automatically reverses the previous steps. This ensures the internal data always matches the external records.
- High-Speed Bidding Engine
The team built a bidding engine focused on extreme speed. They replaced standard libraries with a high-performance framework to reduce memory usage. By using Redis caching and optimized connection tools, Vivasoft enabled the system to process thousands of bids with almost no delay.
- Cloud-Native Task Control
Vivasoft implemented a “leader election” mechanism using the Kubernetes API. This ensures that only one server instance handles background tasks at a time. This setup prevents errors and race conditions when cleaning up data or running scheduled jobs in the cloud.
- Centralized Event Processing
The engineers created a central event machine to manage communication between services. This hub handles over 15 different types of data signals using AWS SQS. It allows the system to respond to real-time changes instantly without putting a heavy load on the main database.
Measurable Results
The implementation of Ad Exchange transformed how Azerion handles its digital advertising operations. These changes led to faster workflows and better performance for their partners.
Single Management Hub
Vivasoft created a central API that connects all external platforms. This allows Azerion to manage all their ad operations from one place. It significantly reduced the manual work needed to sync different networks.
Enhanced Bidding Speed
The engine now responds to bids in under 50 milliseconds. This high speed allows Azerion to participate in more auctions. As a result, they can fill more ad slots and generate more revenue.
Improved System Stability
The platform now recovers from errors automatically. This has lowered the number of technical issues and support tickets. The system stays reliable even when external third-party tools face downtime.
Challenging Features of the Software
The implementation of Ad Exchange transformed how Azerion handles its digital advertising operations. These changes led to faster workflows and better performance for their partners.
Unified Campaign Interface
This feature acts as a bridge between different ad providers. It simplifies complex tasks by letting users update multiple networks with one single action through a simplified dashboard.
Intelligent Traffic Flow
The software separates data reading and writing tasks. This ensures that heavy traffic does not slow down the primary database. It keeps the system responsive even during peak bidding hours.
What's Next?
The project continues to evolve with a focus on integrating additional ad partners and further refining the machine learning algorithms used for content recommendation optimization.
04/04
Azerion.ai: Unified AI Hub
Summary
Vivasoft developed Azerion.ai to simplify how businesses access and deploy artificial intelligence. The platform acts as a central hub where developers connect to different AI models for text, images, and video through one single interface. Vivasoft built a high-speed gateway and an automated marketplace using the Go programming language. This allows users to launch powerful AI tools with one click, while the system handles complex, real-time billing for thousands of users at once.
1+ Years
Artificial Intelligence (SaaS) & Cloud Infrastructure
Multi-Modal AI Orchestration & Cloud Gateway Development
Client’s Business Challenges
Azerion needed to overcome technical barriers to make enterprise AI accessible and financially secure. They had to overcome these specific obstacles:
- High-Concurrency Billing Risks:
Processing thousands of requests at the same millisecond created a danger of “double-spending.” Azerion needed to ensure users could not use more credits than they owned during traffic spikes.
- Fragmented Integration Methods
Developers originally had to use different sets of code for every AI provider. This made building and maintaining new applications slow and complicated.
- Complex Infrastructure Barriers
Most AI tools require advanced server knowledge to set up. Business users wanted to use these applications but lacked the technical staff to manage the underlying cloud systems.
Solutions
Vivasoft engineered a custom cloud-native architecture to solve these problems. The following solutions provided a stable and user-friendly experience:
- Atomic Billing Engine
Vivasoft built a thread-safe system directly into the database. It uses row-level locking to verify and deduct credits instantly. This ensures 100% financial accuracy and prevents any revenue loss.
- Universal API Gateway
The team standardized different AI types into one simple format. Now, developers use a single connection to reach any model. The backend automatically sends the request to the best provider.
- One-Click App Marketplace
Vivasoft created an automated system for launching AI applications. Users can start their own dedicated versions of complex tools without needing any DevOps skills or manual server setup.
Measurable Results
The new platform transformed how Azerion delivers AI services to its global audience. The partnership achieved these measurable outcomes:
Zero Billing Errors
The platform achieved 100% accuracy in financial transactions during heavy stress tests. It maintains perfect credit balances even during high-concurrency usage.
Instant App Launching
The time needed to deploy complex AI tools dropped from hours of manual work to less than two minutes.
Accelerated Development
By unifying five different AI formats into one integration, Vivasoft significantly increased the speed at which developers can build new features.
Core Features of this Software
Azerion.ai provides a complete set of tools for managing and testing artificial intelligence. The platform includes these core features:
Managed Infrastructure Catalog
A marketplace where users can launch and manage dedicated AI tools with customizable hardware and security.
Precision Credit Ledger
A sophisticated system that tracks costs for text, images, and minutes of video in real time with financial-grade accuracy.
Multi-Model Playground
An interactive environment where users can test and compare different AI models side-by-side before using them in production.
Standardized API Interface
The platform uses a “drop-in” design. This allows companies to switch their existing AI tools over to Azerion.ai without rewriting their original code.
What's Next?
The partnership between Vivasoft and Azerion continues to grow as the platform evolves into a comprehensive AI operating layer for businesses. Vivasoft is currently focusing on enhancing the platform’s core capabilities to support larger enterprise requirements.
This involves building more advanced management tools and refining how the system processes complex data workflows. The teams are also working on new ways to further optimize infrastructure costs and improve overall system intelligence.
By focusing on these high-level improvements, Vivasoft ensures that Azerion remains at the forefront of the competitive AI market while providing a secure and scalable environment for its global users.