Projects

Personal Site · 2026

Nik Murano

Applied Intelligence Research · Counter-Deception · AI Defense · Russian Strategic Culture

About

A graduate researcher in Applied Intelligence at Georgetown University. The work connects analytical rigor, AI-mediated reasoning, and the design of democratic intelligence services. The portfolio collects a published research framework under journal review, a defended capstone study, and an active program of research-instrument design across counter-deception, defensive retrieval-augmented generation, and Russian strategic culture — documented under the same evidentiary discipline the work itself argues for.

How Projects Are Organized And Built

The navigation is organized by the kind of work a reader is entering. Project Development explains how ideas move through the Personal OS, review, and source control. Flagship Projects contains the largest public-facing research and manuscript surfaces. AI & Retrieval Security gathers the work on AI-mediated evidence, retrieval poisoning, model extraction, and cognitive-security instruments. Intelligence Methods contains tradecraft, structured analysis, red-team review, bias control, and decision architecture. Security, History & Regions contains the substantive fields and cases: terrorism, Russia, regional and historical studies, comparative genocide, infrastructure defense, reflexive control, cognitive warfare, and counterintelligence policy.

Status
Progress bars are claim ceilings
A high percentage means the project has more of its required artifacts in place. It does not mean the project is finished, externally reviewed, or deployed unless the page says so.
Evidence
Each page says what exists now
The Built and Coming sections separate completed work from planned work, so the page does not imply that a prototype, manuscript, or research instrument is more complete than it is.
Method
Project development comes first
The Development Process page explains the six-stage path behind the whole portfolio: concept, vision, design, build, validate, publish.
Reader Path
Start broad, then enter the instruments
A first-time reader can start with the development method, then move into major projects, AI systems, and substantive analysis with a clearer sense of what is built, what is measured, and what remains unfinished.

How To Read This Site

This page is a professional map of work in different states of maturity. Some projects are defended research, some are manuscript systems, some are working software surfaces, and some are governed plans or handoff packets. The site is therefore organized around claim ceilings: what exists now, what the page can honestly claim, what sources carry the claim, and what remains unresolved.

What Counts As Built
Look for named components
A built claim is backed by visible pieces: routes, rooms, files, ledgers, chapters, datasets, audit outputs, validators, or release packets. If a page describes a concept without those pieces, it is treated as a plan or handoff, not as an operational system.
What Is Not Claimed
Progress is not completion
A high status percentage does not mean external validation, deployment, or publication. It means the project has more of its required public-facing artifacts, source layers, or implementation scaffolds in place.
Source Discipline
Bibliography is not enough
The source sections distinguish representative bibliography from source utilization. The stronger pages say what role each source family plays: method, evidence, counterargument, domain framing, claim ceiling, or future adjudication.
Academic Reader
Start with research design
Begin with the defended capstone, AI rigor detection, Discovery Engine, source-trace layers, and analytical-methodology pages. Those pages best show research design, inference limits, evidence handling, and methodological ambition.
Security Reader
Start with substantive judgment
Begin with the Russia Security Desk, intelligence methods, counter-deception work, structured analysis, and Book Trilogy security arc. Those pages best show strategic judgment, tradecraft awareness, and comfort with uncertainty.
Technical Reader
Start with instruments
Begin with AI and retrieval security, model-assessment pages, source-governance workbenches, and project-development systems. Those pages best show how abstract research gets translated into components, gates, and reviewable systems.

Currently

  • Reframing the Discovery Engine manuscript against the audited claim ceiling, ahead of journal submission.
  • Writing the Cognitive Firewalls capstone monograph, with the inferential ceiling stated explicitly in the methodology.
  • Building the AI Extraction Defense prototype layer by layer — behavioral session classifier first, through forensic attribution chain.
  • Refining the Book Trilogy atlas and chapter-focus surfaces from the May 2026 release package; chapter art now covers all seventy-two chapters, with connoisseurship review and workbench wiring still ahead.

Recent

  • May 2026AI Extraction Defense — five-layer architecture and self-imposed adversarial review pre-published.
  • April 27, 2026Cognitive Firewalls capstone defended.
  • May 2026Book Trilogy iPad handoff packaged the atlas entry, chapter-focus route, quote ledger, rewrite state, and seventy-two chapter-art records.
  • March 6, 2026Quantitative-arm reliability gate met; second outcome variable held at chance, inferential ceiling formally documented.
  • February 18, 2026Capstone Masterwork build pack — full fifty-scenario clean/contaminated workbook corpus preserved as a single coordinated artifact.
  • January 23, 2026Structured-analytic-technique operationalization patch locked for the first dependent variable.

Contact

  • Site nikmurano.com
  • Affiliation Georgetown University · Applied Intelligence