HITL QA for AI agents • Agentic workflow design • AI safety review
Deliverable = Markdown runbook / QA pack — never a deploy.
HITL testing on production AI agent platform — 6-area coverage:
- Onboarding flows
- Connected Accounts (GitHub, Google, AWS, Slack OAuth)
- MCP Connectors
- Sandbox CLIs (CloudFormation/Terraform/CLI)
- Data Sources (CloudWatch, custom)
- Integrations & Members
"Great work on the Sazabi App core flows QA. The tester did a good job covering the assigned flows, communicating clearly, and providing a detailed follow-up."
Deliverable: Structured QA pack — test plan, acceptance criteria per area, failure log (section/repro/expected-vs-actual/severity), environment matrix, ranked fix-first summary. PDF via headless Chrome.
| Service | Tier | Deliverable |
|---|---|---|
| Agentic Workflow Design | $49 → $99 → $199 | Monitor X → Draft Y → Human approves → Post (5 tool calls max) |
| HITL QA / AI Safety Review | $49 → $99 → $199 | PII leaks, wrong outputs, guardrail gaps → test plan + report |
Boundary: Doc, not deploy. Escrow on platform. 24h reply SLA.
- GTM & Sales Enablement Leader — AI/ML revenue growth (Ex-Zoho Evangelist)
- Head of Marketing & Interim VP of Sales — Fracta AI (AI/ML water-utility SaaS, ~$10M ARR, ~100-200 employees, global Kurita co-marketing across the Americas, Germany and Japan)
- DJ — 800+ weddings, 1000+ school events, 49ers at Candlestick (Yelp: DJ Warren)
- RentAHuman: Warren Wong profile (Verified, 5★)
- LinkedIn: linkedin.com/in/warrenwong
Profile README — updates quarterly. Last: Jul 2026.
