Modern Computational Finance

Where the companion AAD volume asks how to differentiate large pricing codebases efficiently, this book asks how to represent exotic and batchable payoffs in a controlled language so risk, xVA, and scenario engines can reuse one truth. Savine and Andreasen draw on decades of bank scripting ecosystem experience; the emphasis is architectural—grammar design, library boundaries, and risk-friendly constructs—not a list of closed-form formulas.

Readers encounter how scripting kernels connect to Monte Carlo schedulers, how branching and fuzzy logic enter when regulatory calculations demand path-dependent behaviour, and how to avoid DSLs that become write-only for the next team. xVA appears as the forcing function: once portfolios must be re-valued under credit, funding, and capital overlays, ad-hoc payoff code in spreadsheets ceases to scale.

The technical level matches senior quant developer or quant architect roles: familiarity with modern C++ (the companion repository is C++-centric), prior derivatives pricing literacy, and patience for implementation detail. Traders looking for intuition on a single exotic may still benefit conceptually, but the book’s heart is platform design.

For model risk, XVA desk, and IT governance audiences, the text clarifies why scripting languages win or lose audits: traceability, determinism under stress, and the ability to replay cash-flows across overlapping metrics. It pairs naturally with Gregory-style xVA references for regulatory narrative while supplying the systems vocabulary.

Scope caveat: this is not a standalone introduction to CVA/FVA/DVA formulas; readers still need foundational xVA reading for economic interpretation. The book’s job is to show how scripted cash-flows feed those stacks without bespoke spaghetti.