Marcos Eduardo Elias
Independent researcher and architect of asymmetric systems
in computation, markets and decision-making.
Some structures work better
when they are not fully visible.
About
I work on problems that sit at the boundary between computation, markets and irreversibility.
My focus is not prediction, speed or scale, but structure:
how systems behave under stress, how information degrades,
and how decision-making changes when symmetry is deliberately broken.
Over the years, this work has taken concrete form in operational systems —
particularly in algorithmic trading, corporate intelligence
and non-anthropocentric approaches to artificial intelligence.
I do not operate as a public intellectual, nor as a conventional consultant.
Most of my work happens in private conversations, closed research loops
and long-term collaborations with a very small number of people.
What follows here is intentionally partial.
Systems
Julia Mandelbrot Trading System
Julia Mandelbrot is an operational computational system for trading and corporate intelligence.
It is not a single model, but an architecture that combines
regime detection, convexity, fractal geometry
and non-anthropocentric decision principles.
The system produces actionable signals
and structural readings of markets and companies,
designed to operate across normal and extreme regimes.
Julia Mandelbrot is in production use
and has been externally evaluated as a high-value computational asset.
Non-Anthropocentric AI (EquiVerse)
Most contemporary AI systems are built around human heuristics,
interpretability narratives and anthropocentric objectives.
My work in non-anthropocentric AI explores
decision systems that remove the human from the center
when human intuition becomes the limiting factor.
These approaches are applied to:
• trading and risk,
• biological and animal systems,
• pattern recognition in non-human environments.
The goal is not imitation of human intelligence,
but structural coherence under uncertainty.
Cryptographic Systems (Mock Data Functions)
I am developing cryptographic primitives
based on structures inspired by Ramanujan’s mock objects.
The focus is not computational hardness alone,
but informational hardness:
systems where local access does not allow reconstruction of the whole.
These designs explore:
• controlled non-holomorphy,
• symmetry that is structurally misleading,
• irreversibility beyond brute-force attacks.
This work is foundational infrastructure,
not a consumer-facing product.
KIAI — Decentralized Exchange Infrastructure
KIAI is a decentralized exchange architecture
designed as a system, not as a marketplace.
It is built around:
• a proprietary oracle,
• autonomous infrastructure,
• architectural latency reduction,
• protection against oracle manipulation and liquidity stress.
Volume is not the primary objective.
Resilience and structural integrity are.
Physical Anchors: Mining
Purely informational systems tend to collapse together.
For that reason, part of my work includes
structured exposure to physical assets,
such as gold mining operations,
treated not as speculation but as systemic anchors.
Notes
On Asymmetry
Not all asymmetries are suggesting exploitation.
Some are necessary for stability.
On Signals
A signal that needs explanation
has usually already expired.
On Intelligence
Removing the human from the loop
is sometimes the most human decision.
On Silence
Some systems degrade
when exposed to excessive attention.
Private
Most of my work happens in private.
From time to time, I engage in
closed conversations, small research circles
and long-term collaborations
with people who operate capital or systems at a high level.
If you believe a private conversation might make sense,
you may leave a brief note below.
No materials are sent automatically.
Not all messages receive a reply.