An AI agent, layered on top of Claude Code, that re‑implements large existing quantitative libraries while preserving their original interface, making adoption seamless, reducing support overhead, and freeing up engineering resources. The agent combines a toolkit of Python scripts, Numba, and external tools with a Claude CLI prompting library to drive dynamic planning‑and‑execution workflows.
This agent is very well suited to very large quant production libraries.
We have drawn on our 15 years of experience as front‑office quants at major investment banks (Morgan Stanley, Citibank, Deutsche Bank, Nomura) to optimize our agent for rewriting quant libraries — which can be somewhat messy — into clean, well‑formed, bug‑free, and extensively documented libraries. These new, cleaned libraries are then very easy to refactor when needed. As a result, valuable resources no longer need to be spent supporting legacy code, and can be reallocated to more revenue‑generating projects.
The agent is language agnostic (C++, Java, C#,…).
One example of the use of this agent is the NumpyPandas project.