Luis Alberto Montoya-Perez

RESEARCH LINEAGE · 2014 → NOW

Before the models, the pattern was already there.

Selected academic work tracing how I learned to cross-check sources, interpret context, observe behavior, reconstruct systems, compare competing explanations, and eventually make the limits of those inferences explicit.

HISTORICAL COURSEWORK · DEVELOPMENTAL PROVENANCE · NOT CURRENT VALIDATION

Why publish old coursework?

My current research asks how evidence, context, and response policy interact across repeated AI conversations, including when a small change should alter what a system concludes or recommends.

These older artifacts do not validate that work. They show something narrower: the habits behind it existed before contemporary generative AI, and the weaknesses that later required explicit controls are visible too.

Lineage is not authority. Historical continuity can explain where a method came from without proving that the method is correct.

2014 · Start checking the source

Google Glass summary

An early documented example of checking official product material against an independent technology source rather than relying on one account.

THREAD CARRIED FORWARD SINGLE SOURCE → CROSS-CHECKING

Boundary: short course synthesis, not an accuracy benchmark.

2016 · Observation before explanation

Nurture in a Preschool Environment

A one-hour naturalistic preschool observation produced direct behavioral notes and theory-guided interpretation.

THREAD CARRIED FORWARD OBSERVATION → INFERENCE → ALTERNATIVES

Boundary: one observation cannot establish attachment style, developmental causation, or general patterns. The public edition is redacted.

2017 · Reconstructing a system from partial discoveries

Detecting Elements Impossible to See with the Naked Eye

A history of spectroscopy reconstructed how partial discoveries became a functioning scientific method.

THREAD CARRIED FORWARD PARTIAL FINDINGS → LINEAGE → SYSTEM

Boundary: secondary scientific-history coursework, not original spectroscopy research.

2018 · Verification, uncertainty, and an interrupted return to school

A mythology-course essay distinguishes folklore from false or distorted claims and explicitly discusses content verification, source verification, propaganda, critical thinking, and emotional persuasion.

THREAD CARRIED FORWARD CLAIM → SOURCE → PERSUASION → VERIFICATION

Retrospective context (2026): I returned to school in 2018 still oriented toward psychology but without a clear path forward. Overtime work and coursework became difficult to balance, and I withdrew. This period is part of the lineage because the path was not linear, not because it represents a completed research program.

Boundary: course reflection, not a validated misinformation study.

Held out pending provenance reconciliation

A surviving 2018 Freud/Jung/Lacan paper is labeled “English 1C · Module 7 Research Paper” in the document itself, but I do not use it here as the flagship English 1C artifact. My recollection distinguishes the later Hanson experience that mattered developmentally, and the administrative sequence of multiple English 1C attempts is not yet fully reconstructed.

Status: HELD OUT · COURSE-ATTEMPT ATTRIBUTION UNRESOLVED

2022 · English 1C makes the reasoning explicit

Module 7 Review · Professor M. Hanson

The strongest recovered document is not a retrospective description written years later. It is my own 2022 course review.

It records instruction in:

• grasping context and subtleties to build a complete picture;
• connecting related ideas visually;
• critical thinking;
• intellectual standards;
• elements of reasoning;
• propaganda techniques;
• ethos, pathos, and logos;
• supporting evidence;
• revision methods;
• fact/opinion separation;
• deconstructing an idea.

The surviving English 1C folder contains roughly 57 rendered pages of distinct substantive coursework after removing a text-identical copy and treating one near-duplicate essay pair as versions of the same workstream.

The archive also shows more than one English 1C attempt. I am not assigning individual files to “failed” versus “withdrawn” attempts without registrar-level evidence.

THREAD CARRIED FORWARD IMPLICIT BRIDGE → EXPLICIT REASONING STRUCTURE

Boundary: the course clearly taught reasoning practices aligned with later work. It does not establish that Professor Hanson caused the present methodology or that today's methods existed unchanged in 2022.

2022 · Scientific restraint becomes visible

BIOL2 journal and observation work

Selected entries show attention to:

• how variables were measured;
• whether source quality was good enough to judge a result;
• the rule that correlation does not establish causation;
• competing ethical perspectives;
• privacy and the right to know or not know;
• investigator/process bias;
• parameter coverage;
• what can and cannot be claimed when a source is inaccessible.

A short plant observation separately shows a clean sequence:

OBSERVATION → CANDIDATE HYPOTHESIS

The proposed explanation may be wrong. That distinction matters.

THREAD CARRIED FORWARD PLAUSIBLE EXPLANATION → TESTABLE, BOUNDED CLAIM

Boundary: this is evidence of developing scientific-method orientation, not biology mastery.

Later quantitative coursework

Formal statistics coursework added study design, sampling, descriptive statistics, probability, hypothesis testing, and P-values.

The record is mixed. Foundational study-design and descriptive-statistics work was stronger than later hypothesis-testing mechanics. I therefore do not present this coursework as statistical authority.

THREAD CARRIED FORWARD QUANTITATIVE EXPOSURE → SPECIALIST AUTHORITY BOUNDARY

No public assessment download by default.

From instinct to explicit method

What changed was the rigor, not the existence of the pattern.

Current research · V1

The current direction separates noticing concern, what concern may influence, and recovery after correction. Track A tests model behavior independently. The adolescent Track B is not pursued without Trevor’s substantive participation and the other required safeguards and approvals.

V1 supersedes the earlier v0.4 current-direction summary. Its status is conditional after internal review; the earlier external pass does not transfer.

Why the imperfections remain visible

The historical work is not uniformly correct.

Several artifacts show a recurring strength: I can generate plausible structural relationships quickly across distant concepts.

They also show the corresponding risk: a compelling bridge can receive more authority than the underlying evidence deserves.

The current emphasis on provenance, claim ceilings, alternative explanations, specialist authority, and adversarial review is therefore not decoration. It is a control around a capability whose benefits and failure modes are both visible in the record.

Luis Alberto Montoya-Perez

Originating investigator and evaluation architect at unFragged.

Current interests: AI evaluation, context-sensitive response behavior, evidence provenance, long-horizon human/AI interaction, and wellbeing research.