Experiment-as-Code Labs: A Declarative Stack for AI-Driven Scientific Discovery

Researchers propose Experiment-as-Code (EaC) Labs, a new paradigm that bridges AI agents with physical laboratory equipment by encoding experiments as declarative configurations compiled to device-level APIs. This framework combines artificial intelligence with automated lab instrumentation through a systems layer that performs safety checks, resource allocation, and job orchestration, enabling AI-driven scientific discovery beyond purely digital environments.

The proposal addresses a fundamental limitation in current AI-for-Science initiatives: the separation between intelligent decision-making systems and physical experimentation. While AI excels at data analysis and simulation, breakthrough discoveries often require real-time observation and adaptation in laboratory settings. The EaC Labs framework attempts to close this gap by treating experiments as software artifacts that can be version-controlled, compiled, and executed across different laboratory environments.

This initiative builds on the growing autonomy movement in laboratory automation, where scientific instruments increasingly expose programmable interfaces. However, previous approaches typically lack the abstraction layer necessary for AI agents to reason about experimental workflows at a high level. By introducing declarative configurations—specifications of what should happen rather than how to make it happen—the researchers enable agents to express scientific intent without detailed knowledge of specific instrument APIs.

The systems layer innovation draws parallels to containerization and orchestration in cloud computing, suggesting that lessons from decades of distributed systems engineering can accelerate autonomous science. Safety checks and resource assignment mechanisms are critical for preventing experimental failures and optimizing laboratory utilization, directly impacting research productivity and cost efficiency.

For the scientific and technology communities, this framework could democratize access to automated experimentation by abstracting away instrument-specific complexity. It potentially enables smaller research institutions to leverage advanced AI-driven discovery without massive infrastructure investments. The science-, lab-, and instrument-agnostic design indicates scalability across academic and industrial research environments.

• →EaC Labs encodes experiments as declarative configurations that bridge AI agents and physical laboratory equipment.
• →The framework combines intelligent hypothesis generation with automated safety checks and resource orchestration.
• →This approach applies distributed systems principles to scientific experimentation, enabling reproducibility and scalability.
• →The abstraction layer allows AI agents to reason about experiments without detailed knowledge of specific instrument APIs.
• →The technology could democratize autonomous experimentation across research institutions of varying scales.