Powering a Natural Language-Enabled Air Force

Kyndi’s current provisioning and deployment framework relies on a network of publicly available tools and services such as GitHub.com, CircleCI, and AWS ECR (public regions), AWS S3 (public regions), to name a few.  Additionally, Kyndi leverages approximately fifty AWS services from the public cloud. Kyndi has confirmed that all but one of these services exist on AWS GovCloud. In order to support installation into a secure environment, Kyndi must conduct research and development to either leverage the AWS and other provisioning and deployment resources in a secure environment or identify and build a novel framework not relying on commonly available deployment productivity tools such as those above.  This effort can be accomplished on an unclassified network to allow security vulnerability assessments as well prior to implementing the resulting stand-alone Kyndi platform that is confirmed to have removed security vulnerabilities on the designated classified network. Once the platform is re-engineered to run independently of the internet and all vulnerabilities are remediated, Kyndi will implement the resulting platform on the designated classified network, ingest classified documents,  tune the model to effectively incorporate DAF-specific terminology and phrases, and then demonstrate the results to the End User and Customers for feedback.  Kyndi will then further tune the model based on the feedback and demonstrate the final version to the End User and Customers for evaluation.