Visual Embassy
Seven years of client work delivered through Visual Embassy, across community stacks — Python, Node, Go, MongoDB — and the Microsoft stack of SQL Server, .NET, and C#. The engagements below are the ones that carry the A.I. and cloud story.
Fannie Mae
Fannie Mae, the Federal National Mortgage Association, supports liquidity in the U.S. mortgage market. Kenneth helped lead a move of Python, Tableau, and Power BI reporting off an on-premises Oracle backend and onto PostgreSQL in AWS. He defined how the Python ecosystem would do the processing, and he prototyped a migration-documentation tool so stakeholders could see progress while the integration was still underway.
FaceFirst
FaceFirst builds facial-recognition technology for security operations. During the COVID years Kenneth worked with data scientists on facemask-compatible recognition, using Python and PyTorch. The front end was Angular 10 and React. Request handling sat on .NET Web API, feeding an analytics dashboard used to identify crime and the people responsible.
Liberty Dental Plan
Liberty Dental Plan is a dental-benefits organization. With its data-science and business-intelligence group, Kenneth built a visual fraud-detection tool and converted Jupyter machine-learning notebooks into something operations could run, aimed at keeping members and cutting fraud. Azure, SSIS, and Blazor supplied pieces of the visualization.
Secure One Financial
He led development of a loan-modification workflow manager and a call-center analytics platform, including the Salesforce dashboards beside them. A later pass put a React interface on MongoDB and a large language model so the center could see agent performance and customer satisfaction, and lose fewer customers.
Toyota Racing Development
Toyota Racing Development is Toyota’s motorsports and performance arm. The team extended Athena, the real-time NASCAR tracking and analytics platform, by moving the backend to AWS Lambda, with Node in the middle and React on the front.
Ingram Micro
Ingram Micro is a global technology distributor. Working with data scientists in the business-intelligence department, Kenneth turned sales and logistics models out of Jupyter notebooks and into production, and guided a data-driven sales-forecasting tool from early sketches to a proof of concept shown to BI leadership.
Puffy
With William Andreozzi, formerly of Weedmaps, he added delivery tracking to Puffy’s social mobile app. The interface was Angular. AWS Lambda and DynamoDB carried a live status and map so a delivery read as a timeline, not a ticket number.