Profile picture of Orhun Oezbek

Orhun Oezbek

Software Engineer -
UBS AG

Orhun Özbek is the GenAI Lead and a software engineer at UBS, specializing in building reusable and scalable GenAI platforms and providing business solutions for banking AI use cases. He leads the UBS RiskLab Model Services GenAI team which develops the RiskLab AI Common Ecosystem platform which streamlines AI development and deployment across UBS through innovative RAG, AI Agents, Benchmarking and Guardrail functionalities.

Orhun will present RiskLab VEGA (Vector Embedding Governance Application) which is a cornerstone of the RiskLab AICE platform. RiskLab VEGA leverages Azure PostgreSQL PGVector to offer governed and self-service Vector Store RAG development and benchmarking capabilities for AI developers. His presentation will include a deep-dive into the challenges of processing complex financial documents and explore advanced indexing and retrieval techniques.

With a background as an ML/AI data scientist, Orhun now utilizes his expertise to build and optimize AI/ML platforms. He is passionate about creating reusable IT systems and advancing AI and machine learning capabilities in the financial sector.

Session Presenting

Tuesday, November 19
5:30 PM - 6:15 PM Greenwich Mean Time
BRK190
The success of GenAI apps is decided by the accuracy of their responses. Using Retrieval Augmented Generation (RAG), you can improve accuracy by grounding GenAI app responses in your data. In this session, explore advanced RAG techniques in Azure Database for PostgreSQL including new vector search algorithms, parameter tuning, hybrid search, semantic ranking, and the GraphRAG approach. See how customers are using these techniques to deploy corporate development platform for GenAI apps.