SystemsEvidenceAccess

Rethink the Obvious.

We craft what holds under load and expose what was destined to fail. No theater, no high-visibility vests.

7
research areas
combined
48+
systems
working together
53k+
data points monitored
275+
hand-coded agents
1
evidence-first ML world model

Areas of Magnification

scaHeECas atjfHnvNBOXhlCD

How complex systems behave: how evidence becomes structure, how relationships change over time, and where one butterfly can move an entire network.

Start Here

turning observations into intelligence

Most organizations have more signals than they can interpret: systems, vendors, markets, events, dependencies, and changes arriving from different directions. We bring them into one evidence-first world model, preserve where every observation came from, and trace what changes through the relationships around it.

What gets lost
What the model preserves
Your most experienced engineer leaves. Six months later, a critical system behaves unexpectedly, and nobody knows why it was designed that way.
Institutional intelligence: operational knowledge, design decisions, dependencies and outcomes remain connected to their original context and provenance.
A supplier in Asia misses one shipment. At the same time, a port is congested and energy prices spike. Each event looks manageable on its own. Production stops three weeks later.
Cross-domain intelligence: supplier activity, logistics, infrastructure, markets and external events are connected to reveal compound risks before they reach operations.
Your board asks why a major investment was approved. The people who made the decision remember the conclusion, but not the evidence that made it compelling.
Decision intelligence: evidence, assumptions, alternatives, conditions and subsequent outcomes remain connected, making the decision reconstructable months or years later.
One of your suppliers quietly starts buying from a cheaper source. Nobody hears about it, until a bad batch shows up and you are the one explaining it to your own customers.
Supply-chain intelligence: multi-tier dependencies, geographic exposure, concentration risk and upstream disruptions are traced through the network to their potential business impact.
A storm shuts down a port on the other side of the world. You do not ship anything through it yourself. Six weeks later your own delivery times start slipping, and nobody can say why.
World intelligence: geopolitical events, trade routes, commodities, infrastructure, suppliers and business dependencies are continuously connected across the external world.
Fuel gets more expensive. Then shipping costs creep up. Then the price on your own invoice quietly rises with it, and nobody remembers when it started.
Causal intelligence: environmental, economic and operational signals are tracked through time so the model can trace how distant events propagate into business conditions.

Causality Chains

A look into Dingir's Mind

Distant events propagate through infrastructure, markets, and decisions. We trace those chains end to end, from the first signal to the effect that actually reaches you.

The Earthquake That Changes a Food Price Months Later
Earthquake
An earthquake is recorded as the initiating event. Its effect is then tracked into local port infrastructure is damaged.
01 / 20

Research Cooperation

built alongside our coop partner

Laura Serna Gaviria directs the Emergent Interaction Lab's research and agent architecture. Lauras Team, Call Laura, and Jarvis all came out of her method. RFI-IRFOS builds what she directs. Her name stays on her work.

Laura Serna Gaviria

Live API

Emergent Interaction Lab · Coop Partner

Research into human-AI interaction since 2023 - the method behind Lauras Team. It is a SWAT architecture: 15 specialised expert agents form the lead team, and together they orchestrate a swarm of 303 specialised sub-agents beneath them. We have tested this directly rather than assumed it: whenever a task could be solved our way or hers, we tried both, and her way has worked in ten out of ten cases so far. That is the actual reason her name stays on everything it produces, not a courtesy credit.

4 of her packages, shown as entry points across engagement phases - the full list depends on where a company is in its process. Full pricing on request via emergent-interaction-lab.fly.dev.

Contact & Disclosures

reach out

One form, even if it's just a first talk: a question, a service inquiry, research collaboration, or a security finding. Findings come to us directly, never through a third-party bug bounty platform. We would refuse to be routed through one ourselves.

Not sure what to pick on the right? General inquiries reaches a human either way - or email one of these directly.

Sending a security finding? See our full disclosure-handling policy - triage, lawful basis, and your choice of credit.

Whatever it's about, expect a reply from a human within 12h of us receiving it.

© 2026 RFI-IRFOS  ·  Graz, Austria  ·  GISA 39261441  ·  UID ATU83405245

Human rights are not subject to negotiation.

Foundation

Intelligence

World Model

Applications

Repositories

Crates (crates.io)

all packages →

this is a useless cookie banner. it's just here to look like one * we don't use cookies, so there's nothing to consent to. don't let anyone tell you otherwise.two buttons, one closes this and throws some confetti. the other literally does nothing.