About

Built on a different
set of questions.

Most AI development begins with a model. Aevaron begins with a question: what would it mean to build intelligence that can be fully understood?

Aevaron is a research-first technology company focused on the design, measurement, and validation of safe, interpretable intelligent systems.

That question shapes everything. Rather than layering products on top of existing black-box systems, Aevaron researches and constructs the underlying cognitive architecture from first principles, deterministic, symbolic, and measurable by design.

The result is a company that operates across two layers simultaneously. At the research layer, Aevaron is developing Synthetic Intelligence, a non-pretrained cognitive architecture whose internal states, transitions, and behaviours can be directly observed and formally evaluated. At the applied layer, Aevaron delivers production AI systems built with the same commitment to transparency and reliability.

The belief is simple: systems that cannot be measured cannot be trusted. Aevaron exists to build the kind of AI that earns trust through engineering, not marketing.

01

Deterministic architecture

Intelligence built from explicit, auditable components. Every internal state, transition, and output has a traceable cause.

02

Measurable stability

Formal metrics track coherence, drift, and behavioural integrity across every version and experimental run. If it cannot be measured, it cannot be validated.

03

Research into production

Foundational research and applied engineering inform each other. Production systems are built on the same rigour that drives the research programme.

04

Long-horizon thinking

Aevaron's work sits at the intersection of cognitive systems research, software engineering, and AI safety, with a focus on properties that matter over time, not just at deployment.

The Founder

Founder

Dylan Jooste

Dylan Jooste founded Aevaron with a conviction that the most important problems in AI are not about scale, they are about structure, observability, and trust.

Self-taught across software engineering, cognitive systems, and applied AI, Dylan brings an independent perspective shaped by building real systems rather than academic abstraction. His work spans both ends of the technical spectrum: from designing the theoretical foundations of the Synthetic Intelligence architecture to deploying production AI systems for commercial clients.

The Synthetic Intelligence project represents years of independent research into what it would mean to build a mind that can be fully understood from the inside. Alongside this, Dylan has shipped document AI systems, intelligent data platforms, and custom AI backends, each informed by the same engineering discipline that drives the research.

Aevaron is the vehicle through which that work becomes a company: a place where rigorous research and practical engineering exist in the same organisation, reinforcing each other rather than existing in isolation.

Approach

Independent research from first principles. No reliance on existing pretrained architectures as a foundation for cognitive work. Every claim about system behaviour is backed by formal metrics and reproducible experiment logs.

Research focus

Deterministic cognitive architecture, stability as a measurable engineering property, behavioural evaluation frameworks, and the conditions under which artificial systems can be said to exhibit genuine coherence over time.

Applied work

Production AI systems delivered across document intelligence, marketing analytics, automated auditing, and conversational agents, all built with explainability and operational reliability as primary constraints.

Location

Cape Town, South Africa. Working with organisations globally.

Work with Aevaron

Whether you are exploring a research collaboration or commissioning an AI system, we would like to hear from you.

Get in touch