Projects
Two threads run through these. The first is finance and data: tools I build to understand how a valuation or a price series actually behaves once you have to implement it rather than read about it. The second is the mathematics underneath, written out by hand rather than called from a library, because implementing something is the fastest way to find out whether I understood it.
Data and finance
Open any of them for the stack and how it is put together.
Numerical methods and econometrics
Smaller pieces, written to understand a method rather than to ship a tool. Each one implements something a library would normally hide, so that the failure modes are visible.