Black-Scholes and friends, packaged as small, easy-to-use Python modules with a real test suite.
from kuantil.pricing import black_scholes_call
call = black_scholes_call(spot=100, strike=100, risk_free=0.05, sigma=0.2, expiry=1.0)
Why
It started during a financial management course in my Master of Management, with the thought: is it fun if I turn this into a module?
It was. It is also the most reliable way I know to find out whether I actually understood a formula. A derivation can be nodded along to, but a function has to produce the right number, and a test has to agree. Several things I was sure I understood did not survive that.
This is the newest thing here, and it points somewhere different from the rest: quantitative finance sitting next to the ML and data work.