Security / Automation / Systems

Farhan Shafee.

Security Engineering ×
AI Automation × Fintech Systems

I build security tools and automation systems, with a background in fintech software engineering, test automation, and backend financial systems.

Explore my work GitHub LinkedIn

01 / Selected work

Selected work.

All work

Security engineering / AI systems

AegisGraph

Active

A security investigation engineering platform for synthetic fintech telemetry: scenario replay, evidence-grounded AI, hypothesis tracking, and detection regression.

The model is not the system of record. Typed claims require validated evidence; persistent review decisions remain human-controlled.

AegisGraph V2 synthetic Atlas replay paused with three alerts, the first correlated incident, and causal processing stages. — Synthetic Atlas replay connects observations to alerts and correlation, with compressed source-time gaps disclosed. Recorded in the September 23 public-mode rehearsal.

Fintech investigation / AI systems

CreditOps AI

Completed

Recorded investigations of synthetic trading incidents, connecting diagnostic evidence with RFQ/order impact, assessment review, and human handoffs.

Recorded playback makes no fresh model or MCP diagnostic calls; accepted assessments remain advisory and require human review.

Fintech systems / Test engineering

TradeFlow

Completed

A synthetic trading workstation that makes test ownership, API contracts, and three-browser workflow validation inspectable.

Test-owned sessions connect exact-record UI assertions with API state checks.

TradeFlow synthetic order blotter with server-side filters, lifecycle states, and a LIMIT order ticket. — Synthetic order book and LIMIT ticket. Each session owns its records; submission creates a NEW order without simulating execution.

Financial systems / Quality engineering

CreditFlow QE

Active

A transactional credit-card simulator with independent Java, Cucumber and REST Assured automation, plus executed CI and browser-test evidence.

Independent Java/Cucumber automation validates a synthetic transactional credit-card API.

CreditFlow QE hosted simulator after a synthetic $50 payment — Actual Playwright journey: $200 outstanding, $800 available and one committed payment.

Security engineering: application, detection, remediation & cloud

02 / Writing

Notes from the work.

Sep 17, 2026~17 min read
Building an Evidence-Grounded AI Security Investigation System
The engineering decisions behind AegisGraph: deterministic incident scope, bounded AI claims, server-side evidence validation, and human-controlled case decisions.

Writing archive RSS

03 / Engineering notebook

Questions under investigation.

Detection signals, test isolation, and human review in AI-assisted workflows. Exploratory work, with scope kept explicit.

Open the notebook