Discovery Loop plans AI systems that automate experimental cycles, starting with machine learning research and engineering

Discovery Loop says it is building systems that automate the full scientific experimentation cycle: proposing tests, running evaluations, analyzing results, and iterating. The company will first target machine learning workflows, using its own tools internally before expanding to broader science and engineering challenges.

The pitch reflects a growing push toward AI-driven R&D, where models and compute infrastructure compress research timelines by running many experiments in parallel. If effective, this approach could shift scientific and engineering work from manual iteration toward scalable, automated discovery loops.