Our client operated a Connected Worker Platform that helped manufacturing companies digitize work instructions and production processes. However, years of accumulated technical debt slowed the platform down instead of optimizing workflows and limited its ability to scale.
Hence, to support the client’s goals and strengthen the platform for enterprise use, we modernized its architecture, improved performance, and reduced API latency by more than 94%.
The Starting Point
Even if it seemed like the platform had been working fine, once you scratch the surface, all the technical debt that was left behind by multiple development teams has emerged:
- API response times of 5+ seconds on critical endpoints, which feels like eternity when a factory worker is waiting on a production line;
- No SSO integration, making enterprise onboarding a manual, friction-heavy process;
- There was no public API for third-party integrations;
- Customer onboarding required manual migration of large document libraries;
As the client expanded its customer base and attracted larger enterprise organizations, these limitations became a significant business risk and threatened the platform’s ability to scale.
Priorities Behind the Build
The key priorities included:
- Improving platform stability and performance
- Enable enterprise integrations through SSO and a public API
- Automating manual onboarding processes with AI
- Improving overall code quality and maintainability.
How It was Built
- Team Structure
The project involved one of our Senior Backend Developer working directly within the client’s product team alongside a Product Manager, Tech Lead, Frontend Developers (Angular), and Mobile Developers (Capacitor).
- Development Process
Instead of planning the whole process upfront, development followed an Agile approach with two-week sprints and capacity-based planning.
To accelerate delivery, QA responsibilities were shared within the development team, and to minimize risk, new features were first released to 1-2 pilot customers before wider deployment.
- Communication Process
During the project, our developer wasn’t working as an external contractor, but within the client’s team, communicating through Jira and Slack, they joined sprint planning sessions and regular reviews.
The embedded model ensured complete alignment between all stakeholders.
Key Solutions We Delivered
- Enterprise SSO with OpenID Connect
We implemented full OpenID Connect support, allowing enterprise clients to connect their existing identity providers, such as Microsoft, Google, or any OIDC-compatible system directly to the platform.
Consequently, enterprise onboarding became significantly faster and no user migration was required when onboarding new corporate clients
- Backend Performance Optimization
Through systematic profiling at both the database and Node.js runtime level, we identified and eliminated the bottlenecks dragging down API performance.
As a result, overall latency decreased by more than 94%, reducing critical endpoint response times from over 5 seconds to just 0.3 seconds.
- Versioned Public API for Integrations
Using Fastify with full versioning and strict TypeScript schemas, we built a stable public API layer for enterprise integrations.
This solution not only simplified B2B integrations, but also ensured that future internal API changes would not disrupt existing customer integrations.
- AI-Powered Document Translation
Using Microsoft Azure AI services, we automated the conversion of physical and digital work instructions, including PDFs, scanned documents, and legacy SOPs, into the platform’s native format.
This significantly reduced the manual effort required to onboard customers with large document libraries.
Results and Impact
The modernization initiative delivered measurable business and technical outcomes:
- API response time reduced from 5+ seconds to 0.3 seconds.
- Enterprise SSO introduced with zero user migration requirements.
- Launching public API, enabled enterprise integrations for the first time.
- Through AI-powered document processing, manual onboarding work was reduced.
Final Thoughts
Technical debt doesn’t just slow down development, it also affects the relationship between a business and its customers.
By combining deep backend expertise, an embedded team model, and a clear focus on the highest-impact problems, we helped transform a platform that was struggling under its own weight into one ready to scale with enterprise demands.
If your business is experiencing the similar challenges, contact us.
Headquarter: