AAD training The AAD course taken by quant teams at leading banks.
AAD itself, not one tool: the principles, the adjoint maths and how to write adjoints by hand, then how to apply them in a large code base with any AAD library. Led by Jörg Lotze, co-founder and CTO, who built XAD and presented it at CppCon 2026; on site or remote, with a hands-on workshop on your own code. It ends with a review of your code, 30 days of questions and a certificate.
From the course to production
AAD · Training
First AAD pricer checked against bumping in six weeks
Reported by the client · Bank quant team
From AAD training to AAD in production
After the AAD course, the bank's own quants made their first pricer AAD-enabled with XAD and checked it against bumping within six weeks; it went into production about four months later, after validation.
Full case studyThen AAD enablement on your own code: a fixed-scope build with XAD, the code yours. AAD enablement.
Adjoint algorithmic differentiation (AAD)
2 days · Quants, quant developers
- Sensitivities and the need for speed: SIMM and MVA, FRTB SA, SA-CVA, XVA hedging
- Writing adjoints by hand: from the maths to code, line by line
- Finite differences, tangent and adjoint modes, tapes and operator overloading
- First and higher order; control flow
- AD in large code bases: discontinuities, external libraries, payoff scripting, complex numbers
- Memory: checkpointing, trade-offs, tape serialisation
- Performance: parallelism and thread safety, manual AD for bottlenecks, mixed modes
- Advanced: pathwise and vector-mode Monte Carlo, American Monte Carlo, implicit functions, testing strategy
- Worked examples, with XAD as the reference tool: Monte Carlo and PDE pricers, SIMM, FRTB SA, SA-CVA, CVA sensitivities
On site or remote, for one team or several. Worked examples run on XAD, and the workshop runs on your own code.
“Six weeks after the course our own people had the first AAD pricer matching bumping. It went into production about four months later, once validation had signed it off.” Head of Quantitative Development, Bank
Related
- ExplainerWhat is AAD (adjoint algorithmic differentiation)?The idea in plain words, with a worked example.
- XADXAD, the AAD library for C++ and PythonAutomatic differentiation for pricing and risk code, open source or by subscription.
- ServiceAAD enablementAAD built into your own pricing and risk code with XAD.
Book AAD training for your team
The AAD course taken by quant teams at leading banks, taught by the team that built XAD, with a hands-on workshop on your own code.