Our mission

Advancing Implicit Intelligence Research

We are working on advancing AI research in the fields of implicit learning and implicit reasoning to build systems capable of transferring knowledge to other domains through experience compression and related approaches. Our goal is to pave the way toward enabling machines to understand implicit knowledge about the world

Explore Our Vision

Simulating the experience compression node in machine learning

Contact setaleur.aplamda@gmail.com

Research AI Implicit · SAI

Focus Structural Intelligence

Setaleur Aplamda

We are enabling the next AI revolution, by building a new breed of AI systems that (1) extract structural knowledge instead of memorizing statistics, (2) compress experience into dense, reusable, transferable structure, (3) commit to bold structural hypotheses instead of hiding behind soft probability distributions, (4) know the exact boundary of their own knowledge.

Our main goal is to build intelligent systems whose intelligence scales with the density of experience they compress, not with the number of parameters they memorize.

The dominant paradigm learns by minimizing a scalar loss over a fixed topology, distributing prediction credit across millions of weights that never commit to a structural hypothesis about the world. Confronted with the unfamiliar, these systems still normalize their uncertainty into a confident answer because softmax always sums to one, whether the input is known or not.

Setaleur Aplamda is developing AI Implicit, a gradient-free architectural family that reorganizes its own connectivity topology the way a learning brain does: forming geometric prototypes from experience, measuring true structural distance rather than dot-product similarity, and refining that structure under reconstruction pressure rather than prediction accuracy.

Structural transfer lets these systems recognize the same relational pattern across domains that look nothing alike on the surface. Epistemic transparency lets them abstain the moment an input falls outside everything they have ever learned a capability closed-form softmax normalizers are architecturally incapable of.

Setaleur Aplamda is building this Structuralist AI (SAI) paradigm and its architectural family Deep Transducers, TSNet, Linear Networks and advancing applications where reliability, controllability, and safety are not optional: industrial process control, automation, robotics, wearable devices, healthcare, and beyond.

Founded and led by Momen Ghazouani, Setaleur Aplamda is a research programme built on a falsifiable thesis, published in the open, from day one.

We share one belief: intelligence is not prediction. It is structure and the topology is the knowledge.

We can build the future of trustworthy AI together: with industry partners, product developers, and the global research community, through open publications and open source.

If this vision resonates with you, come join us:

Contact Hiring Press Updates