Ravenrock designs AI infrastructure and operates the ventures that run on it. Four businesses. One stack. One founder.
Ravenrock builds AI infrastructure and runs businesses on top of it. The founder is the architect and the operator. No gap between design and execution.
Multilingual macro research, market education, and AI-assisted analysis. Macro, vol, rates, FX, commodities, digital assets. EN / ZH / JP markets.
Research on the platform is authored by Dorian personally — not by Ravenrock. Professional-grade research built for practitioners who act on their own conviction.
An internal research program exploring systematic macro strategies. Concept and dashboard prototype stage. Not currently accepting capital, advisory clients, or accredited investors.
Independent AI consulting, courses, and tools for founders, teams, and individuals.
ZMACRO's verification architecture extended to threat intelligence and adversarial AI.
On-chain capital markets infrastructure. Structured derivatives, tokenised macro strategies, and advisory for operators building at the intersection of TradFi and DeFi.
Hands-on guidance for individuals and teams adopting AI into their workflow. Practical, not theoretical.
Courses and structured training on AI tools, workflow design, and applied research methods.
Independent research engagements and advisory on AI system design. Scoped to the actual problem.
Co-development or advisory arrangements. One active slot per quarter.
One person. Four ventures. The same person who designed the architecture built the code and runs the businesses.
Self-taught across macro trading, full-stack engineering, and AI system design. Operating across US, Japan, and Chinese-speaking markets.
Research and consulting operate independently under the Dorian personal brand — separate from Ravenrock as an entity.
For partnerships, research collaborations, or anything that doesn't fit a standard session.
Pick a time. Talk directly with the founder. Three formats available.
A working philosophy at the intersection of reflexive markets, irreducible uncertainty, and artificial intelligence. Not a framework to trade with — a way of thinking about why frameworks break.
None of these are mine. What might be mine is the insistence on holding all three at once — because each one, alone, produces a different kind of overconfidence.
Standard economics assumes markets converge toward a true value that exists independently of what anyone believes. Reflexivity rejects this entirely. Participants' beliefs shape prices, and prices shape participants' beliefs — a feedback loop, not a mirror.
Operationally, this means every model — including those built into ZTrader.AI — is itself a participant in the system it describes. Publish a signal widely enough, and the signal changes what it was measuring. That's not a flaw to engineer around. It's the actual texture of markets.
A black swan isn't just rare — it's an event the model assigned near-zero probability to because the training data came from a world before it happened. Every backtest bets the future resembles the past closely enough to matter. The times it doesn't are the only times that matter.
The response isn't predicting the unpredictable — that's a category error. It's designing structures that survive being wrong: capped downside, optional upside, deep suspicion of any system whose value depends on a calm, mean-reverting world.
Most AI applied to markets chases the wrong target: better prediction. But reflexivity says the system is unpredictable in principle, and black swan theory says the tail events that matter most are by definition outside the training distribution.
The honest target is epistemic infrastructure — systems that compress information faster than a human can, flag when a claim is unverified, and explicitly track where the model stops applying. That is the actual design principle behind ZMACRO's verification layer: not "what will happen," but "what do we actually know, and where does our framework break down."
A signal published widely enough becomes reflexive — the act of measuring changes what is being measured. Build for that, not against it.
Black swan theory says the edges dominate history. The model should know precisely where it fails, not only where it works.
Not "what will happen" — but "what do we actually know with what confidence, and where does this framework stop applying."
Macro research and AI-assisted analysis built on the premise that markets are reflexive and tail risk is the main event, not the exception. Every thesis includes its own falsification criteria.
Visit ZTrader.AI →Claim extraction, source weighting, explicit data-state flags — surfaces what's known vs assumed, not forecasts with false confidence. Verification as design principle.
View Architecture →The entity that exists to build infrastructure shaped by this philosophy — across financial intelligence first, and any domain where the same epistemics apply.
Visit Ravenrock →For teams building AI systems for uncertain domains — not just markets — who want the reflexivity and tail-risk lens applied to their own architecture and decision systems.
Book a Session →The goal was never to predict the market. It was to build something honest enough to admit what it doesn't know — and fast enough to be useful anyway.
Direct access to the founder. AI architecture, system design, workflow strategy, or a general advisory session. Pick a type, pick a time.
First contact. We figure out if there's a fit. No commitment, no preparation required.
Architecture review, system design, workflow audit, or strategic advisory. Come with a specific problem.
Working session. We build, design, or map something together in real time. Bring your stack, your data, your codebase.
For complex engagements, partnerships, or anything that doesn't fit a standard session. We'll respond within 48 hours.
How we handle your data. What we collect, what we don't, and why.
Read →The rules of engagement for all Ravenrock services and products.
Read →Financial and regulatory disclosures. Built for independent operators who make their own calls. Not a registered fund.
Read →How to reschedule or cancel a booked session, and our policy on no-shows.
Read →Last updated: June 2026. Ravenrock Fintech LLC ("Ravenrock") is committed to protecting your privacy.
Solely to respond to your inquiry, confirm and manage your booking, and communicate about your engagement with Ravenrock. We do not contact you for marketing purposes without explicit consent.
Inquiry and booking data is retained for up to 24 months or deleted upon request. To request deletion, contact us directly via the inquiry form.
Ravenrock Fintech LLC is incorporated in California, USA. Data is handled in accordance with applicable US law. Users in the EU/EEA may have additional rights under GDPR.
Last updated: June 2026. By accessing Ravenrock's website or engaging with its services, you agree to the following terms.
Ravenrock provides AI infrastructure consulting, system design advisory, and access to products including ZTrader.AI. All services are provided as-is, subject to availability and the founder's discretion.
Ravenrock's advisory services deliver operator-grade intelligence — for principals who act on their own conviction, at their own risk. Ravenrock is not liable for decisions made based on information provided during sessions or through its platforms.
These terms are governed by the laws of the State of California, USA. Any disputes shall be resolved in the courts of California.
Last updated: June 2026.
Nothing published by Ravenrock Fintech LLC, ZTrader.AI, or Dorian (in any capacity) constitutes investment advice, financial advice, or a solicitation to buy or sell any financial instrument. All research and analysis is for informational and educational purposes only.
Ravenrock Fintech LLC is not a registered investment advisor, broker-dealer, or fund manager. Ravenrock Capital is a pre-launch vehicle and is not currently accepting capital. Any future fund operations will be subject to applicable regulatory compliance.
Past analytical accuracy does not guarantee future performance. Markets are inherently unpredictable. Users of ZTrader.AI and related products assume full responsibility for their own trading and investment decisions.
Ravenrock does not accept affiliate fees or undisclosed compensation for product recommendations. Any partnerships or commercial relationships will be disclosed explicitly.
Last updated: June 2026.
One reschedule per booking allowed, provided it is requested more than 24 hours before the original session time. Rescheduled sessions are subject to available slots.
Failure to attend a booked session without prior notice will be recorded. Repeated no-shows may result in restricted access to future bookings.
Use the link in your booking confirmation email, or contact us directly via the inquiry form on this page. Include your booking reference and preferred new time if rescheduling.
A small, considered collection for the desk and the field. Notebooks, instruments, and a few quiet artifacts — designed first, manufactured once they earn it.
Each one earned its place by solving something specific. If it didn't make the desk quieter or the read clearer, it isn't here.
A markets edition for chart sketches, trade theses, and the notes that don't fit a spreadsheet. Hybrid grid pages, lay-flat binding that survives a real working session.
A build edition for system maps, architecture sketches, and the diagrams that happen before the code does. Same binding, dot-grid pages tuned for structure.
Machined from aluminum, finished matte black with a brass cap band. Weighted for sessions that run long. Refillable with a standard gel cartridge — built once, used for years.
The raven mark, reduced to its load-bearing geometry and cast solid in blackened steel. Small enough for a pocket, heavy enough to notice every time you reach for your keys.
A 52-card deck across checklists, risk prompts, and mental models. Pull one before a session, or shuffle when the thinking gets stale. Letterpress finish on matte stock.
A solid brass weight with hand-finished facets. No moving parts, no charge cable, no app. It holds the desk down, and gives the hand something to do while the mind works.
Solid bronze, 38mm, serial-numbered. The first run is the only batch sold individually — after launch, it's issued, not purchased.
The notebook, the pen, and the Signal Cards deck — together, at a small saving. The minimum kit for a working desk.
Beyond the first seven, a few more pieces are taking shape. Shown here as concepts, not commitments.
Every name under consideration, organized by series. Most of it is just a name on a page right now.
"Most of this list will never ship. That's by design — a long list is easy to write and hard to stand behind. The first seven are the test: if they hold up in someone's hands, the rest of the catalog earns the right to exist."
No store yet — just a list. When the first issue is ready to ship, this is who hears first.
Restricted to mandate clients and research partners. All sessions are logged.