Fractional and interim CTO/CIO leadership for FinTech, payments, and AI in regulated finance.
We are brought in for a leadership gap or a decision nobody can afford to get wrong: a departing CTO, a payments integration, a sponsor bank or BaaS program under review, a compliance deadline, or machine learning that works in a notebook and not in production.
Work covers payments and real-time rails (FedNow and RTP), core platform and ledger architecture, PCI DSS and SOC 2 readiness, BaaS and sponsor-bank integration, and getting models into production with governance a regulator will accept. Based in Toronto, remote-first across North America.
azdv.co · What we do · Case studies · Book a call
Two categories, and it is worth being clear which is which.
Current work.
- payment-rail-reference - dated, sourced, machine-readable parameters for US payment rails. FedNow, RTP, ACH and Fedwire, with every value carrying its source URL and the date it was checked.
- finstatement-parser - structured JSON extraction from US bank and credit-card statement PDFs. A reference implementation, not a supported product, and not published to PyPI.
Archived. Everything else here predates the current practice, most of it by a decade. It is archived rather than deleted because deleting history to look newer is its own kind of dishonesty. It is not representative of what we do now and it is not maintained.
Anything we put here is a reference implementation or a maintained reference document. It is published so the approach can be read and reused, not because we are running a community project. Issues and pull requests get reviewed periodically rather than on a schedule, and that is stated on each repository rather than left for you to discover.
If something here is wrong, a correction with a failing example or a source citation is the fastest thing to get merged.
Longer technical pieces live on the site rather than in repositories: