I work across finance and product in digital assets. The code here comes out of questions my day job raises and does not answer.
- QuantLite: a fat-tail-native quantitative finance toolkit for Python. Extreme value theory in the core risk calculations, Hidden Markov regime detection, regime-aware portfolio construction, and backtest forensics. On PyPI, MIT licensed, actively maintained.
- cartulary: a draft standard for portable receipt chains in agent-initiated payments. Append-only hashed receipts verifiable by recomputation, with a TypeScript SDK and adversarial test cases for tampered, truncated, and partially-disclosed bundles. Apache-2.0.
- FolderSync: a small cross-platform directory sync tool, standard library only.
- Older repositories are archived and left public. They are from 2016 and are best read that way.
Elsewhere: prasant.net · LinkedIn



