
Dan Barrett
CEO and founder
Background in fintech and building production AI systems.

OpenProse is an open-source agent framework for event-driven, multi-agent workflows.
It helps teams encode institutional knowledge in-house and define requirements for auditing and guardrails.
An introduction to OpenProse and its local and cloud options, followed by a guided demonstration of a financial workflow.
We’ll email you links and materials from our AI Engineer NYC workshop when they’re available.
Model documentation, regulatory reporting and compliance checks repeat on a cycle, follow required formats and depend on specialist knowledge.
This work takes time from the people who build and maintain the models.
OpenProse captures the requirements in a plain Markdown contract. Agents follow the contract and return their work with evidence for review.
The example documents a USD SOFR OIS discount curve—a financial model used to value future payments—using public data and a fictional bank’s conventions.
Agents prepare the documentation from three layers:
Shared model-documentation requirements.
The bank’s template, terminology and evidence requirements.
The model’s technical brief, sources and worked example.
Information that only the institution can provide is left for its team to supply.
This is a methodology demonstration, not a regulatory submission or a claim of compliance.
Model documentation is one example. The same approach can apply to regulatory reporting, verification checklists and questions about written procedures.

CEO and founder
Background in fintech and building production AI systems.

Founding AI Engineer
PhD chemist, entrepreneur and independent researcher.

Founding Engineer
Built software at Google and Nuzzel; co-founded Newslit.
Start with a single regulatory, compliance or reporting workflow your team repeats.