Volume VII / Sanctions Intelligence
FinanceCase
Miami Software Journal
A fintech engineering case study about API design, data normalization, compliance workflows, and explainable decision support.
Case Study
Finance /
Compliance
Case Study
This is a case study showing how I turned public U.S. Treasury sanctions data into a documented API, then used Subjective Logic in a web application for making decisions and evaluating risk.
A fintech engineering case study showing how public OFAC sanctions data can become a documented developer API, then a Python decision-support prototype for risk screening.
This project has two layers: SDN OpenAPI is the TypeScript serverless API layer, and ED 209 is the Python FastAPI decision layer.
The Problem
Messy entity data needs explanation.
Financial and compliance systems often need to reason about names, aliases, countries, sanctions programs, and missing context. A fuzzy score alone does not explain whether a match should be blocked, ignored, escalated, or sent back for more information.
The Build
Part 1 / API Layer
SDN OpenAPI
The TypeScript serverless API layer. It exposes U.S. Treasury OFAC SDN sanctions data through OpenAPI, Swagger UI, ReDoc, fuzzy search, entity lookup, metadata endpoints, and refresh-ready data handling.
Part 2 / Risk Engine
ED 209
The Python FastAPI decision layer. It consumes sanctions search results and turns them into explainable screening decisions, surfacing evidence, uncertainty, missing information, and a recommended next action: clear, block, escalate, or gather more information.
Architecture
From source data to decision support
The point is not just querying OFAC. The point is showing how public compliance data can move from normalized source data, to a developer API, to decision support that explains uncertainty instead of hiding it.
- 01
Treasury OFAC SDN source data
- 02
TypeScript OpenAPI layer
- 03
Fuzzy search and entity lookup
- 04
Subjective Logic evidence model
- 05
Decision: clear, block, escalate, or gather more
Why It Matters
A binary flag is often too crude.
A useful screening system should explain what evidence exists, what evidence is missing, and whether the next action is clear, escalate, block, or gather more information.
This page frames the work as a software engineering artifact: API design, data normalization, compliance workflows, and explainable decision support.
I can turn messy public financial and compliance data into documented APIs, then build reasoning systems on top of it.
Technical Stack
What I would build next
- Batch screening for multiple entities.
- Country risk dashboard built from sanctions counts and entity types.
- News ingestion with source citations.
- Historical sentiment snapshots.
- A market hypothesis page clearly labeled as speculative analysis.
Recruiter Summary
I built a finance and compliance case study showing how public OFAC sanctions data can become a documented TypeScript API, then a Python FastAPI decision-support prototype. SDN OpenAPI handles the developer API layer. ED 209 consumes sanctions search results and uses Subjective Logic to surface evidence, uncertainty, missing information, and next actions instead of returning a brittle match/no-match result.