This paper addresses the need for novel software system development (SSD) practices in finance. It proposes Empirical Modelling as a novel approach for SSD in finance. This approach aims at finding a suitable framework for studying both the traditional and the emerging computing culture to SSD in finance. First, the paper studies the change in the financial industry and identifies key issues of the application of computer-based technology in finance. These key issues are framed in a wider agenda for computing in finance. Second, the paper motivates a paradigm shift at the computational level to meet the wider agenda for computing in finance and overviews the distinctive qualities of model building in EM that are of particular relevance to this wider agenda and that can potentially support a paradigm shift at the computational level. Third, the paper considers four case studies that reveal how EM technology can provide a framework for SSD in finance: in the financial enterprise, in the financial market, for financial engineering, and for financial analysis. The paper concludes with the need for new paradigms for SSD adaptable to domain specific needs.