ZRT Vercel AI Systems Suite
Resume Evidence RAG Auditor
DEMO JSON RETRIEVAL

Audit resume claims against project evidence before tailoring bullets.

Editable job description, local retrieval corpus, claim verification states, unverified-claim flags, evidence gaps, and grounded bullet export.

Plain English

This tool checks whether resume claims are backed by real project evidence, then helps rewrite only the claims that can be supported.

DEMO JSON RETRIEVALClaim verificationEvidence gapsGrounded bullets

Readiness Console ReadinessA quick view of how many claims are supported and whether the resume needs review.

Readiness ReadinessREADY means all entered claims have evidence. REVIEW means some claims need proof or rewriting.REVIEW
Verified score Verified scoreThe percent of claims that matched project evidence.67%
Verified Verified claimsClaims that found a matching project evidence record.2/3
Evidence records Evidence recordsThe local proof documents the demo searches through.4

Matched job terms: evals, agent

Inputs InputsPaste the job description and the resume claims you want checked.

Evidence Corpus Evidence corpusThe small library of project proof this demo searches. In a larger system, this could be a database or document store.

Vercel Build Doctor Agent

Verified app classifies deployment failures, redacts secrets, exports incident reports, and runs 8 eval fixtures.

vercel | next.js | log parsing | redaction | evals | reports

Enterprise Agent Workflow Studio

Models mock tools, agent graph generation, approval gates, deterministic evals, and audit exports.

agents | approval gates | audit | tool registry | evals

AI Gateway Failover Playground

Simulates unified chat routing with outage fallback, cost/latency telemetry, request traces, and eval fixtures.

gateway | provider routing | fallback | latency | cost | trace

Resume Evidence RAG Auditor

Audits resume claims against local evidence, flags unsupported claims, and generates grounded bullets.

rag | retrieval | claim verification | job matching | grounded bullets | evals

Claim Review Claim reviewEach claim is marked VERIFIED when evidence supports it, or UNVERIFIED when it needs proof.

VERIFIED

Built Vercel Build Doctor Agent with log parsing, redaction, structured reports, and eval harness.

Confidence 95% | Job overlap: evals, agent

Evidence: Vercel Build Doctor Agent | vercel, log parsing, redaction, evals, reports

Evidence: Resume Evidence RAG Auditor | rag, evals

Evidence: Enterprise Agent Workflow Studio | evals

VERIFIED

Built enterprise-safe agent workflow demos with approval gates and audit reports.

Confidence 91% | Job overlap: evals, agent

Evidence: Enterprise Agent Workflow Studio | approval gates, audit, evals

Evidence: Vercel Build Doctor Agent | evals, reports

Evidence: Resume Evidence RAG Auditor | rag, evals

UNVERIFIED

Led production AI platform for millions of users.

Confidence 28% | Job overlap: evals, agent

Evidence: Resume Evidence RAG Auditor | rag, evals

Evidence: Vercel Build Doctor Agent | evals

Evidence: Enterprise Agent Workflow Studio | evals

Scale claim requires direct usage evidence.

Production platform claim exceeds verified demo/project evidence.

Claim lacks matching local evidence or appears inflated beyond verified project scope.

Cross-App Build Repair Workflow Export reportThe final handoff packages verified evidence and report text without inflating claims.

STEP 1 | Build Doctor
Paste build logs

Secret-redacted log input

STEP 2 | Build Doctor
Diagnose root cause

Root cause, confidence, evidence lines

STEP 3 | AI Gateway
Show trace

Trace timeline with status labels

STEP 4 | Enterprise Studio
Suggest patch

Patch checklist and audit gate

STEP 5 | Resume Auditor
Export report

Markdown report for reviewer handoff

Evidence Gaps + Checklist Evidence gapA missing proof point. The safest action is to remove the claim, rewrite it, or attach evidence.

Evidence gap

Led production AI platform for millions of users.

Add a project URL, commit, test output, deployment URL, or work-log entry before using this claim.

Remove or rewrite unverified claims.

Attach deployment URLs and commit evidence to verified claims.

Run eval fixtures before exporting role-specific bullets.

Keep demo-only claims labeled accurately.

Grounded Bullet Export Grounded bulletA resume bullet generated only from supported claims, not inflated claims.

- Built Vercel Build Doctor Agent with log parsing, redaction, structured reports, and eval harness; evidence: Vercel Build Doctor Agent, Resume Evidence RAG Auditor, Enterprise Agent Workflow Studio; aligned terms: evals, agent.
- Built enterprise-safe agent workflow demos with approval gates and audit reports; evidence: Enterprise Agent Workflow Studio, Vercel Build Doctor Agent, Resume Evidence RAG Auditor; aligned terms: evals, agent.

Markdown Report Export Markdown reportA copyable report for reviewer handoff, generated from the current audited claims.

# Resume Evidence RAG Audit Report

## Summary
2/3 claims verified against local evidence.

## Claim Review
- VERIFIED: Built Vercel Build Doctor Agent with log parsing, redaction, structured reports, and eval harness.
- VERIFIED: Built enterprise-safe agent workflow demos with approval gates and audit reports.
- UNVERIFIED: Led production AI platform for millions of users.

## Tailored Bullets
- Built Vercel Build Doctor Agent with log parsing, redaction, structured reports, and eval harness; evidence: Vercel Build Doctor Agent, Resume Evidence RAG Auditor, Enterprise Agent Workflow Studio; aligned terms: evals, agent.
- Built enterprise-safe agent workflow demos with approval gates and audit reports; evidence: Enterprise Agent Workflow Studio, Vercel Build Doctor Agent, Resume Evidence RAG Auditor; aligned terms: evals, agent.

## Safety Notes
- The MVP uses JSON mock retrieval, not a live vector database.
- Unverified claims are flagged instead of rewritten as facts.
- Generated bullets are grounded only in matched evidence records.