owen crabbe.Get in touch

Owen Crabbe / Applied AI & research engineering

Applied AI for real estate research and everyday decisions.

I build tools that turn complex information into source-linked briefs, explicit scenarios, and practical next steps. My work spans real estate diligence, reusable reasoning methods, and evidence-aware fitness planning.

See the projects Explore the writing

Anderson, Indiana · Open to relocation.

AI-generated editorial illustration
Working artifacts. Visible assumptions.A portfolio with a source trail.

01Selected projects

From question to working system.

Explore the implementation,
the decisions, and the limits.

Real estate researchPublic implementation

Cividian
Site Diligence

A public, source-linked diligence pipeline that turns a site question into an investigation brief. Missing evidence stays visible, and zoning remains subject to planning-office confirmation.

Evidence retrievalScenario analysisCitation checks

Evidence: public source and a reviewed evaluation of 16 scripted cases. It is not a live factual-accuracy or investment-outcome benchmark.

Read the case study Inspect the source Explore the demo
AI-generated editorial illustration

Planning & reflection / Working product

Phroneme

A browser-local activity journal, seven reusable skills, and a reviewed evidence contract, published as the open-source MindForge-Skills toolkit. Plan a week, record what happened, and keep source relationships and uncertainty in view.

Evidence: public journal and validator code, plus 54 software tests at the linked revision. Software tests do not establish reasoning quality, adherence, or health outcomes.

Read the case study Try Fitness Lab Inspect the toolkit
How to read the work

Case studies connect implementation to evidence. A software test, a model evaluation, and a real-world outcome answer different questions.

Read about verification boundaries

02Publications & technical writing

Notes from the build.

Self-published analysis.
September 2026.

These engineering notes explain the work and its limits. They are self-published technical writing, not peer-reviewed papers.

Follow future notebook entries.

Subscribe via RSS

03Research in progress

Questions
worth testing.

Proposed experiments on data quality, source support, and the behavior of AI systems. Each direction names a method and an intended output.

Explore the research agenda Proposed studies. No published results or preprints listed.
  1. 01

    Civic data curation & retrieval

    How do deduplication and provenance checks change source retrieval and grounded answers?

    Proposed
  2. 02

    Evidence-bounded answers

    Do claim-level citations and explicit unknown states improve the usefulness of an answer?

    Proposed
  3. 03

    Temporal evidence & abstention

    Can better treatment of dates prevent a system from relying on a superseded rule?

    Proposed

04About Owen

Useful tools.
Inspectable work.
Better questions.

I’m Owen Crabbe. I build applied AI tools for real estate research and everyday decisions. Cividian Site Diligence and Phroneme show how that work becomes software. My background includes commercial real estate work with CENTURY 21 Bradley Realty, Inc.

My research interests include AI for real estate and economics, cognition and reasoning methods, and evidence-aware fitness. Studies on this site stay labeled proposed until their methods and results are published.

I use AI-assisted development and research workflows. The case studies link implementation details to public artifacts, while the notebook records engineering decisions and open questions.

Explore my public repositories

Research, engineering & collaboration

Start with
a useful question.

Interested in the projects, an applied AI workflow, or a research collaboration? Share the question you’re working through.

Start a conversation owencrabbe@owencrabbe.comOpens your email app.