Marco Lewis

Enterprise AI Transformation Lead · AI Systems Architect

Operating Layers of Intelligence · NTWX-STUDIO

Location: Phuket, Thailand — remote / hybrid; open to Singapore employers and relocation once mutual interest and package are clear. Does not claim Singapore or Australia work rights.

Email: mark@ml-nightworx.io

LinkedIn: linkedin.com/in/ml-ntwx

Site: marco-lewis-v3.pages.dev · Live proof: DOME · Hawk

Summary

Enterprise AI Transformation Lead and AI Systems Architect who turns commercial constraint into a governed working system — problem framing and outcome logic, operating-model design (roles, workflows, decision rights), adoption design, and accountable go-live so value does not die in the pilot.

I sit at the product–engineering–commercial boundary: align executives, operators, and builders through ambiguity; connect AI / cloud / data choices to business outcomes; and keep authority separate from model spend. Live proof includes 7ARCH (governed multi-agent delivery method — decision rights, HITL, go-live gates), DOME + Claire (solo-built customer-acquisition / agentic commerce), and Hawk → NIGHTHAWK (AI-search discovery into a multi-surface product). Practice work runs through ML-Nightworx / NTWX-STUDIO as the studio and proof environment; Marco Lewis is the hire face.

~30 years of commercial operations (media → luxury marketing → global trade/supply → agentic systems) underpin commercial acumen for investment framing and stakeholder leadership. Based in Phuket; remote / hybrid; open to Singapore employers and relocation once mutual interest and package are clear. Does not claim Singapore or Australia work rights.

Selected proof

1. 7ARCH — Governed AI delivery method / operating system

Deep-link: https://marco-lewis-v3.pages.dev/ (Work / Approach)
Role: Designed and operates 7ARCH as the delivery operating system for governed multi-agent / AI programmes — discovery through adoption and go-live.

Decision rights & operating model: authority routing separate from model spend; locked roster / surface seat rules — patterns for governance, roles, and readiness without inventing named enterprise deployments.
Commander Firewall / HITL: human approval gates before irreversible commercial actions; outbox stays DRAFT until approve — change control that keeps velocity honest.
Reflector go-live gates: structured GO-LIVE ALLOWED | BLOCKED checks (incl. retrieval quality, provision dry-run, sovereignty) before production release — benefits realisation discipline, not demo theatre.
Method and governance IP — not SaaS and not a named enterprise deployment. ADE-F.T is a formalisation name for authority vs compute lanes, not a product SKU.

2. DOME + Claire — Customer acquisition / agentic commerce (solo-built)

Live: https://dome.ml-nightworx.io/ · Role: Solo conceptualised, designed, built, and manages.

Commercial problem → working customer-facing agentic system: catalogue intelligence, guided agent (Claire), SEO→AEO / agent-native discovery, qualified retailer handoff.
Production path on Cloud Run with HITL on edit/audit; early analytics only — evidence of implementation under constraint, not scaled ARR claim.
TypeScript/Node primary; Python-capable; GCP evidenced surfaces.

Evidence honesty: Live surfaces and early GA4 — not invented client logos, utilisation, or booked benefits.

3. Hawk → NIGHTHAWK — AI-search discovery to multi-surface product (light)

Live: https://hawk-aeo.ml-nightworx.io/ · Role: Solo-built acquisition → product pattern.

Free AI-search / AEO discovery surface that hands off into a multi-channel defence and aerospace intelligence product (web glass, API, Telegram) on Cloudflare.
Stack evidence: Pages + Workers + KV; GDELT feeds; Gemini analysis; Stripe checkout; USDC agent micropayments with spending-policy enforcement.
No ranking, traffic, indexing, or commercial-performance claims — cite as shipped acquisition→product architecture, not growth metrics.

Earlier career (condensed)

~30 years commercial operations across media, luxury marketing, global trade/supply networks (Europe / Middle East / Australia — design, import/export, SWIFT, logistics, wholesale/retail), and later relocatable housing / flood-mitigation projects. Continuous adaptation under live P&L conditions — commercial acumen for investment framing and stakeholder alignment. Trade ops kept quiet here; furniture branding not led.

2024 – Present: Founder, ML-Nightworx / NTWX-STUDIO — agentic products, AgentOps, and governed delivery IP (flagships above). Studio and proof environment for reusable patterns; hire face remains Marco Lewis.

Capabilities

Domain	Evidence-backed strengths
Problem framing	Commercial constraint → outcome logic → investment case without invented metrics
Governed adoption	Decision rights; HITL / HALT; roles & surfaces; go-live gates; adoption design
Hands-on build	Solo end-to-end agentic systems (DOME+Claire; Hawk→NIGHTHAWK) — not slide-deck-only
Executive ↔ eng	Align leadership, operators, and builders through ambiguity at the product–eng–commercial boundary
Cloud path	GCP (Cloud Run, Firestore, Vertex/Gemini) and Cloudflare (Pages, Workers, KV) evidenced surfaces
Commercial acumen	~30 yrs live P&L history; benefits framing without invented ARR or headcount theatre

Education

Computer Science — Institute of Technology, Ultimo, Sydney NSW

Location

Phuket, Thailand — remote / hybrid. Open to Singapore employers and to relocation once mutual interest and package are clear. Does not claim Singapore or Australia work rights.