Projects

Ten data systems that ran, or still run, in production: continuous ingestion, accounting reconciliation, failure recovery and predictive modeling. The corporate cases appear anonymized for confidentiality, with no company, client or counterparty names, and no financial volumes. What remains is the architecture decision, the trade-off and the mechanism holding the result up. The one exception is the LCK Spring 2024 dataset, public on Kaggle and auditable by anyone.

Financial and banking data

Five ingestion and reconciliation pipelines over transactional, banking and fiscal data.

Digital assets and blockchain

Three integrations across an exchange, institutional custody and direct blockchain node reads.

Open data and machine learning

Predictive modeling in production, and a public dataset anyone can verify.