Tudor Serban

Working on the Cresta Knowledge Assist RAG pipeline with Python and Elasticsearch. Researching retrieval strategy improvements for increased relevance and how LLM judges deal with reasoning about relevance.

Working on the Document Understanding offering from UiPath using .NET framework / .NET core. Responsible for creating and maintaining several workflow foundation (WF) activities and associated algorithms for OCR, document classification and document data extraction. In this role I have invented (patent granted) a vector search based document classification and splitting algorithm that is used in the Intelligent Keyword Classifier activity in UiPath Studio.

Native video syndication platform that connects publishers and advertisers. Working as part of the API team with .NET Core 2.1.

Healthcare industry application using ASP.NET Core WebApi 1.1, Entity Framework, AngularJS. I am the technical lead on the project and supervise and mentor the other developers. Paid special attention to non-functional characteristics such as unit testing, code coverage, code analysis and security in order to deliver true quality to the customer.

Working as part of the XenDesktop development team. To date I have contributed bug fixes and features to 8 different XenDesktop releases. Currently involved in Citrix Workspace Cloud. Enabled the transition to the Cloud and supported the Cloud DevOps team to ensure customer satisfaction.

Working as part of the XenDesktop development team. Took ownership of several components and shared SDKs used by the XenDesktop control plane and contributed with feature development, bug fixes and support for consumers. Promoted to Software Engineer II in April 2014.

The internship was about learning and applying WPF and C#.