Building Neuro Symbolic AI models. We are inherently traceable (not a blackbox). Many orders of magnitude cheaper. Our models are capable of long term planning & reasoning unlike incumbents.
Current Status
Raised $800,000.
Completed our alpha model and working on our beta model.
In advance trial conversations with 2 prospects
Problem or Opportunity
Current generation AI models are a blackbox. Their cost of training increases quadratically with the size of the data. Their business model with the size of the models hinders them to address enterprise data privacy concerns. Due to the API only approach, enterprises lose their IP over time.
Solution (product or service)
* Our architecture solves all the challenges with current generation transformer based LLMs.
a) We are inherently traceable (not a blackbox)
b) Order of magnitude cheaper.
c) Our models are capable of long term planning & reasoning unlike incumbents.
d) Our business model/delivery addresses the enterprise data privacy concerns.
e) Enterprises can build their own IP by licensing our domain specific models
Business model
We deliver domain specific AI models for enterprise customers. Our platform allows our enterprise customers to train/fine tune using their data and that fine tuned model becomes their IP over long term. We charge licensing fee and annual maintenance of our platform.