We are a team of engineers and tech leaders who developed Domain-Aware Generative AI as a way to free enterprise teams from laborious, tedious tasks – giving them more time to develop better solutions. Rohirrim originated with one late-night, all-hands Zoom meeting ahead of an RFP deadline. Determined to make those scrambles a thing of the past, we built a Generative AI platform that is trained solely on internal, unstructured data. Now known as Rohan, our platform is used by global enterprises to enhance critical business processes and accelerate growth.
Working with some of the largest volumes of unstructured data in the world, a pivotal moment shifts our vision to a new paradigm. We realize automated tagging and document AI can drastically improve how data is normalized and used to train Domain-Aware Generative AI.
Early alpha prototypes and the foundational technology called “generative stitching”, which accelerates domain-Aware Generative AI, leads to excitement showing potential of how these novel techniques work in cooperation with LLMs.
With foundational technology stabilized, Rohirrim is formed to solve near and dear pain points and deliver enterprise value using generative AI. While still in stealth mode, early customers validate the technology and product fit.
With initial seed funding, and early customers, RohanRFP is released in a closed Beta. A partner ecosystem develops, and tangible value is created without compromising security. Proposal efficiency benchmarks show a 2,461-4,137% efficiency gain when customers introduce RohanRFP into their proposal workflows.
Rohirrim continues its enterprise journey developing research and prototypes around unsupervised natural language inference. Using document AI and document understanding techniques alongside word embeddings, unsupervised training for a hyper domain-aware 200B parameter LLM begins to take shape.
