Calculation
A machine computes.
Independent experimental research
Expand what humans can perceive, understand, and do.
Experimental research on designed representations for reasoning in complex systems.
We are investigating whether deliberately structured representations can make important computational structure usable for human reasoning and intervention.
Current evidenceTechnical + synthetic
Human validationNot yet begun
01 / The question
What if computational structure could become usable by humans themselves?
Complex systems can contain structure that is mathematically meaningful but difficult for a person to perceive or use directly.
Aukemir is testing whether deliberately designed representations can make some of that structure operational for human reasoning — not simply easier to view, but potentially usable when making intervention decisions in unfamiliar systems.
Can a learned representation support reasoning beyond what was explicitly taught?
02 / Why it matters
A better interface is not necessarily a new capability.
Modern computational systems can calculate, predict, visualize and recommend with extraordinary power. Aukemir asks a different question: can computational structure be transformed into something a human can learn, reuse and transfer?
Representation
A human sees structured information.
Transfer
A human uses learned structure somewhere new.
The experiment exists to determine whether the third is actually happening.
03 / Current program
Transferable intervention reasoning
The first Aukemir research program studies whether a deliberately structured representation can support intervention decisions across dynamic systems different from those encountered during training.
The experiment is designed to discriminate genuine transfer from simpler explanations.
Memorization
Can performance be explained by remembering trained cases?
Shortcut learning
Are superficial cues sufficient to explain success?
Conventional visualization
Does a strong conventional interface perform equally well?
Explicit instruction
Can the same result be produced by simply teaching the rule?
Arbitrary coding
Does the specific representation matter?
Machine recommendation
Is human representation unnecessary because a machine can supply the answer directly?
04 / Evidence
What exists. What does not.
Aukemir separates technical construction from external reproduction, human evidence, workflow evidence and commercial evidence.
Research hypothesisActive & falsifiable
Internal technical architectureBuilt
Internal synthetic evaluationCompleted
Held-out technical evaluationCompleted internally
External computational reproductionPending
Human testingNot begun
Human learningNot demonstrated
Far transferNot demonstrated
New functional human capabilityNot demonstrated
Biological workflow valueNot demonstrated
Commercial validationOpen
Internal technical success is not human evidence.Open Evidence Registry
05 / Build record
Built as a sequence of questions, not a sequence of claims.
01
Scientific question
What would have to be true?
02
Falsifiable thesis
What evidence would discriminate the hypothesis?
03
Prior-art challenge
Has the underlying idea already been solved?
04
Formal specification
Define the first technical object precisely.
05
Calibration
Make the computational system numerically trustworthy.
06
Representation architecture
Translate structure into a controlled human-facing form.
07
Experimental architecture
Separate transfer from shortcuts and alternative explanations.
08
Synthetic integration
Exercise the system end-to-end before human execution.
09
Held-out technical evaluation
Test against unseen technical conditions.
10
External review
Next
11
Human validation
Pending
06 / Long horizon
The first experiment is deliberately narrow. The question behind it is not.
Can functional human capabilities themselves become objects of rigorous design?
That possibility has not been demonstrated.
07 / Founder
Isabella Salcedo Tuiran
Founder & Research Lead
Isabella is a medical student, researcher and scientific founder working across biology, complex systems and experimental deep-tech research.
Her broader scientific work spans molecular cancer biology, epigenetics, transcriptomic interpretation and biomedical research. At Aukemir, she leads the scientific thesis, experimental architecture and research program.
Treat human limits as hypotheses.
08 / Contact
Criticism is useful.
Specific criticism, scientific discussion, and potential research collaboration are welcome.
isabella@aukemir.com