L3:automotive-fraud-dashboard · AI Wing · Listed

Fraud review dashboard and security remediation

A client dashboard delivered in two days, followed by remediation against the client's full security assessment.

Client: Mitsubishi Motors (MMKSI)
Role: Primary dashboard engineer
Period: Mar 2026 to Oct 2026

Results

2 days
Time to the first working client dashboard
Based on repository history; development started from a team prototype.
Verified
~38 findings
Assessment findings addressed through 16 specs in under two weeks
About half were closed at the last recorded update. The re-assessment result is not public; no passing result is claimed.
Verified
A+
Mozilla Observatory security-headers grade
Measured after CSP and security-header hardening.
Verified

Problem

Mitsubishi Motors (MMKSI) needed a working fraud review dashboard quickly, starting from a team prototype. The next requirement was remediation against the client's security checklist.

What I did

I developed the team prototype into a working dashboard with a web tier, FastAPI, Document Intelligence and two-pass LLM review with live progress.
I led security-assessment remediation through 16 specs, covering a deny-by-default web tier, HMAC-signed backend calls, revocable sessions, Prompt Shields, database managed identity and security headers.

Tech stack

Next.js
FastAPI
Document Intelligence
Azure AI Foundry
Cosmos DB
Blob Storage
App Service
Azure AI Content Safety