Airbnb’s Open-Source GraphQL Framework with Adam Miskiewicz - Software Engineering Daily Recap

Podcast: Software Engineering Daily

Published: 2026-02-05

Duration: 56 min

Summary

In this episode, Adam Miskiewicz discusses Airbnb's open-source GraphQL framework, Viaduct, which centralizes data access, reduces operational overhead, and enhances developer experience. He shares insights into its creation, architectural principles, and the future of backend development in the context of AI.

What Happened

Adam Miskiewicz, a principal software engineer at Airbnb, takes listeners on a journey through his career and the evolution of Viaduct. He highlights his non-traditional path into the tech industry, transitioning from small companies to a major player like Airbnb. Over his nearly eight years at Airbnb, he witnessed its growth from a 500-person engineering organization to a robust 3,000-member team. This backdrop sets the stage for the discussion on how Viaduct emerged amid the company's shift from monolithic architecture to microservices.

Viaduct was born out of a need for a unified data layer as Airbnb began adopting GraphQL and moving away from its Ruby on Rails monolith. Adam describes being pulled into the GraphQL working group shortly after joining Airbnb, where engineers from various disciplines collaborated to explore GraphQL's capabilities in the context of their new microservice architecture. He emphasizes the importance of centralizing schema and simplifying functionality to avoid fragmentation and duplication in data access patterns, ultimately enhancing the developer experience and performance of the platform.

Key Insights

Key Questions Answered

What challenges does Airbnb face with microservices?

As Airbnb transitioned to a microservices architecture, they encountered challenges such as fragmented ecosystems, scattered data access patterns, and duplicated business logic. Adam explains that these issues can lead to an uneven developer experience, which is detrimental to productivity and system efficiency. To combat this, a unified data graph with a consistent execution layer was seen as necessary to centralize schema and simplify functionality across teams.

How did Viaduct originate at Airbnb?

Viaduct's inception occurred shortly after Adam joined Airbnb when he was involved in a working group focused on GraphQL adoption. This group consisted of engineers from various teams who were tasked with figuring out how to integrate GraphQL into their growing microservice architecture. Adam recounts that they needed to rethink their presentation services as they moved away from their Ruby on Rails monolith, which led to a more structured approach to service development within Airbnb.

What architectural principles guided the development of Viaduct?

The architectural principles behind Viaduct were influenced by Airbnb's shift from a monolithic structure to microservices, emphasizing a clear separation of data services, presentation services, and derived data services. Adam notes that this structure facilitates better management of the services and allows for more efficient data retrieval methods, ultimately contributing to a more seamless integration of GraphQL within their ecosystem.

Why did Airbnb decide to open-source Viaduct?

While the episode does not delve deeply into the specifics of Airbnb's decision to open-source Viaduct, Adam mentions the collaborative nature of the working group and the broader engineering community's need for robust solutions in managing APIs and data schemas. Open-sourcing Viaduct likely reflects Airbnb's commitment to contributing to the tech community and fostering innovation in data management practices.

What is the future of backend development according to Adam Miskiewicz?

Adam expresses that the future of backend development will be significantly shaped by AI technologies, particularly in the realm of unified data layers. He suggests that as systems evolve, the integration of AI could enhance how data is managed and accessed within backend systems, leading to more sophisticated and efficient engineering practices that can adapt to the rapidly changing landscape of technology.