Creative AI Systems Builder
I build the systems, bridges, workflows, and applications that let new companies outrun traditional businesses — better ideas, shipped as real tech, not decks.
Includes AI-as-a-service platforms and the workflows that sit on them.
ML-Nightworx. Solutions for problems not yet identified. Stay ahead of the curve.
For recruiters and hiring teams Ask Marco for an executive hiring summary, role-fit view, or portfolio evidence brief.
Capability domains
Three domains, each tied to a filed case. No borrowed logos, no testimonials, no invented metrics.
Multi-agent work stays accountable—unsafe decisions and surprise invoices stay off the table.
I define decision rights, approval paths, safety pauses, go-live checks, and cost controls so AI systems remain accountable as they scale.
A consult becomes a qualified retailer handoff — CAC work, not chatbot theatre.
I design the catalogue, the consult, and the retailer introduction so a conversation becomes a qualified handoff — with the limits of the system visible.
Humans and agents find an honest teaser, then a live product they can actually use.
I separate free discovery from the product itself: teasers stay honest, the handoff is attributed, and the live system is something a buyer can actually use.
I do not sell a “power-user” persona. I build governed systems that work under commercial pressure.
Operating principles
Personal virtues, written as commercial operating lines. The case studies are the proof; this is the filter.
Loading principles…
Selected work
Each case opens with Status, Role, Evidence, and Not claimed. Live versus partial stays explicit. No invented metrics, no borrowed logos, no fake testimonials.
Qualified mattress-retail handoff from a sleep consult — catalogue, discovery, and a concierge that can be audited. Solo-built. Early GA4 only.
The operating system I run deliveries on. Authority routing is not model routing. ADE-F.T names the formalisation — it is not a SaaS SKU.
Defence and aerospace intelligence: free discovery that hands off to a web interface, API, and Telegram bot.
The three case studies show depth; these live builds show current range.
A small selection of active product experiments and system surfaces. Supporting evidence of hands-on range—not client endorsements, investment advice, regulated financial services, or scaled commercial claims.
Crypto macro intelligence
Live market-intelligence interface (BTC/ETH/SOL/DeFi/macro) with Hermes-led query and voice.
For informational and research purposes only. Not investment, financial, or trading advice.
onos.ml-nightworx.ioPublic AI-search / infrastructure-discovery surface.
No ranking, traffic, or indexing performance claims.
infra-aeo.ml-nightworx.ioPayments/workflow experiment for controlled digital and machine-to-machine commerce patterns.
Experimental. Not a claim of verified live payroll transactions.
payflow.ml-nightworx.ioIndependent ML-Nightworx builds. Live status, feature availability, market data, and integrations may change as the systems evolve.
How I work
Thirty years through news, luxury retail, global trade operations, and housing — then ML-Nightworx agentic systems. The method is the same: name the real constraint, gate irreversible actions, ship something that can be audited.
Name the real commercial problem before choosing a model.
Who may decide is agreed before spend is routed.
If a required fact is not in hand, we stop. A pause is a successful outcome.
Commercial sends, payments, and go-live stay in draft until a person approves.