Citi · 2025–26

Data Controls Scanner

Agentic AI for compliance discovery

An agentic AI pipeline that scans codebases for data elements and controls, replacing manual compliance audits.

  • Python
  • FastAPI
  • Agentic AI
  • LLMs
  • RBAC
FastAPI services
90+
RBAC review workflow
Multi-role
replacing manual audits
Automated

The problem

Compliance needed a map of the data elements and controls that live across many repositories. Building it by hand meant long, repetitive audits that couldn't keep pace with change.

What I built

  • An agentic AI pipeline that reads codebases and identifies data elements and controls
  • A metadata scanner that filters and batches files so large repositories stay tractable
  • 90+ FastAPI services behind the platform
  • Multi-role access control (RBAC) for reviewers, owners and admins

How it works

The agents do the reading and the humans make the decisions. Findings flow into role-based review queues, so every result is owned and auditable.

Built at Citi. Descriptions stay at the level of problem, capability and outcome. No internal data, algorithms or client details are shared.

Want to talk about systems like this?

Email me