With the rise of ab-initio computations, promising new materials are being predicted at an unprecedented rate. Yet, the synthesis of these materials remains challenging, often requiring months or even years of trial-and-error experimentation. In this talk, I will discuss the development of an autonomous laboratory for inorganic materials synthesis, also known as the A-Lab. This platform leverages recent advances in text mining and machine learning to extract synthesis heuristics from the literature, which are used to design experimental procedures for materials predicted by high throughput using DFT calculations. Bespoke robotics carry out the proposed experimental procedures, resulting in samples that are characterized with X-ray diffraction and electron microscopy. Machine learning comes into play once again to interpret the characterization data and decide whether the experiments were successful in forming the desired material. In cases where the initial synthesis attempts are not successful, a theory-driven decision-making algorithm instructs the robots on which experiments to perform next. To demonstrate the A-Lab’s utility, I will discuss how it was used to autonomously synthesize 40 different materials over the course of just three weeks, as well as the lessons learned from successes and failures in the corresponding experiments.