Scores of startups have hit the scene in recent years with tools that promise to reduce the complexity of building artificial intelligence software. Alegion Inc., an Austin-based contender, believes that the key to easing the challenge lies in AI itself.

The startup today announced that it has raised a $12 million funding round led by RHS Investments to advance its vision. Alegion provides a platform that helps companies streamline the labor-intensive task of putting together training data for their machine learning projects.

Teaching an AI to perform a task like identifying animals in a picture requires feeding the model massive volumes of information. Human workers must assemble a collection of animal photos, draw a box around every specimen so the AI knows what to focus on and add tags containing each species’ name. In large projects, there may be need for as many as hundreds of thousands of photos or other training files with multiple descriptor tags each, which makes the data assembly phase incredibly time-consuming.

Alegion aims to take the burden off AI teams’ shoulders. Companies can upload their AI training datasets to the startup’s platform and have each item labeled, as well as annotated if a project so requires, by human specialists. Organizations can draw on Alegion’s network of thousands of contract workers or bring their own staffers. 

Alegion provides AI-assisted annotation tools that help workers organize datasets faster than they normally could. In crowded photos, for example, a highlighting algorithm can automatically select all the visible objects to save time. A second set of algorithms assesses the accuracy of the labels and notes that workers add to determine if a redo may be in order.

“Just as assembly lines incorporate power tools and robotics to enable scale, ML model development will require machines training machines to achieve the highest levels of model confidence,” Alegion Chief Executive Nathaniel Gates said in a statement.

Alegion says that its platform lends itself to most enterprise AI projects. Companies can use the platform to create image datasets for computer vision projects, text or audio collections if they’re building a natural-language processing algorithm and video sets when they need a mix of visual and auditory data. 

Despite competition from a number of better-funded startups, Alegion has managed to establish a serious market presence. Its customer base includes major brands such as Walmart Inc., Condé Nast Inc, Airbnb Inc. and Home Depot Co. 

Photo: Unsplash