Current lab automation solutions require extensive programming by specialized engineers, making them difficult to adapt to the high variety of experiments common in scientific research. This limits the effectiveness of automation for many tasks scientists face, hindering reproducibility and efficiency. Scientists often communicate experimental intent implicitly, a nuance that existing robotic systems cannot understand or translate into explicit actions.

Tetsuwan is developing AI agents that combine natural language processing with laboratory robotics to enable autonomous execution of complex wet lab experiments. The system decodes scientists' high-level instructions and translates them into actionable steps for lab robots, bridging the gap between experimental intent and robotic execution. By augmenting lab robots with automated calibration, liquid class characterization, and improved collision detection, Tetsuwan enables scientists to reliably use these robots for complex, high-differentiation tasks. This technology aims to create AI scientists that can assist human scientists by conducting more reproducible experiments, freeing them to focus on data analysis and hypothesis development, and ultimately automating the scientific method.

Natural language processing to decode scientific intent from high-level instructions.

Two-way translation between experimental intent and robot execution using linear programming and rules-based systems.

Automated calibration and liquid class characterization for lab robots.

Improved collision detection for reliable autonomous operation.

The primary target audience includes scientists and researchers in life sciences, particularly those in cancer therapeutics, precision medicine, and drug discovery, who seek to improve the reproducibility and efficiency of their experiments.