Skip to content

ROLE_ALIGNMENT.md

Latest commit

 

History

History
24 lines (19 loc) · 1.34 KB

ROLE_ALIGNMENT.md

File metadata and controls

24 lines (19 loc) · 1.34 KB
Raw

ROLE_ALIGNMENT — AgentPayOps

The control plane for agentic spending: as AI agents start paying for things autonomously (x402), AgentPayOps reviews each payment against policy and routes approve / escalate / block — with an audit trail and a human in the loop. It never moves money on its own.

Forward Deployed AI Engineer ★ (Palantir, Scale AI, OpenAI, Salesforce)

  • Business-critical, messy operational problem (autonomous spend governance) with production auditability and a policy engine measured by evals — npm run evals reports a confusion matrix + falseApproveRate (a risky payment wrongly approved) that must read 0.
  • Guardrail posture: the agent decides and routes; humans approve. Money never moves autonomously past the threshold.

AI Solutions Architect (OpenAI, Deloitte, AWS)

  • Policy-as-data (policies with category caps, allow/deny vendors, approval thresholds), x402 payment integration, and an audit log — governance architecture, not a prompt.

AI Product Engineer (Productboard, Databricks)

  • Full-stack Next.js + Supabase + Gemini, with a pure, testable decision core (evaluatePayment).

Proof

  • npm run evals → EVAL_SUMMARY.md (accuracy + falseApproveRate over labeled scenarios, driving the real src/lib/policy-engine.ts). Mirrors the Litmus harness (github.com/anilandcode/litmus).