Decision infrastructure
Product state, evidence, claim ceilings, handoff and bounded movement.
Founder / Product Architect / Independent Researcher
Logan Davis
Founder & Product Architect of Kingan Logic, building pattern-intelligence and decision systems for real-world environments where evidence, state, authority, correction and action cannot safely be collapsed into one answer. The work connects software architecture, AI evaluation, human–AI continuity and domain-specific operating systems without treating model output as automatic decision authority.
01 / CURRENT WORK
Operator first. Researcher by the problems that followed.
The products grew from a recurring operating problem: important signals are usually messy, distributed across people and systems, and easy to flatten too early. My work has centered on finding the pattern, preserving the evidence around it, and turning that signal into a repeatable operating structure without pretending uncertainty has disappeared.
Human–AI continuity
Preserving the human's working map, corrections, source state, permissions and unresolved branches across long workflows.
Applied intelligence
Multifamily, physical assets, logistics and disaster-response systems where the same evidence/state discipline is forced to survive different environments.
Controlled AI research
Evaluator validity, benchmark failure, repair, decision quality, human reconstruction burden and model-assisted reasoning under explicit claim ceilings.
02 / CURRENT RESEARCH
Testing the mechanism, including when it fails.
Kingan Logic separates product validation from scientific efficacy. Controlled studies, blind evaluations, evaluator research, naturalistic failure records and cross-lane experiments are preserved even when results are null, mixed or interrupted. Two public research records are now live on Zenodo, and the broader paper program examines permission, evaluator failure, transfer, human reconstruction burden, peripheral capture and protective-gate validity.
03 / SYSTEMS
Current product families.
The public system pages describe the problem, current implementation surface, evidence state and current claim ceiling without exposing protected mechanics.
SYSTEM / 01
Kingan Logical State
Evidence-backed product-state control for what ran, what applies, what may be claimed and what survives whole-product handoff.
890/890 hosted deterministic checks
SYSTEM / 02
Human Signal & State Preservation
Human-owned map and state preservation across long AI-assisted workflows, with bounded transfer and receiver nonownership.
Human-owned state + bounded handoff
SYSTEM / 03
AI Integrity, Repair & Reentry
Execution truth, evidence/permission separation, signed custody and bounded repair/reentry.
Signed OCI + confidential runtime
SYSTEM / 04
Asset Management Engine
Multifamily decision intelligence connecting evidence, variance, authorization, execution and outcomes across the asset lifecycle.
RC2 100% statement/branch coverage
SYSTEM / 05
Alternative Asset Account
A protected live governed account for physical assets, evidence, identity, custody, rights, market state, value and disposition.
Protected staging live
SYSTEM / 06
Predictive Logistics
Contradiction-aware physical route reconstruction, address-specific patterns and outcome feedback.
351 Python tests + 7 browser flows
SYSTEM / 07
Disaster Mapping & Hazard Response
Live official-source ingestion with hazard, consequence, response feasibility, responder danger and action authority kept separate.
Live NWS/AWS + 600-case evaluator
04 / ROOTS
The patterns came first. The software came later.
I did not come to this through computer science. I came to it after more than two decades of operating work where the job was often to notice what the pro forma, the monthly report or the system had not explained yet.
Birmingham, Alabama
I built what I needed.
In multifamily and asset management, that meant reading the gap between plan and reality—rent, capital, ownership history, physical condition and operating sequence—and knowing when a handful of small signals added up to a different answer. Later, I found myself doing the same thing with a package: the route, hub behavior and a missed cutoff told me it was going to be late before the tracking language could explain why. The pattern was there; the system could not hold it.
Then I bought smart glasses, took them into a consignment shop with my nephew, held up a pair of shorts and asked what they should cost. The answer was essentially that they could not tell me. That was the line for me. I did not need another AI that could describe an object or repeat a status. I needed one that could hold the map—the evidence, branches, corrections and context—while I decided what mattered next. Kingan Logic grew from that pressure. Today I build from Birmingham, after a career spent operating across very different markets. The setting changed; the habit did not: notice the pattern early, preserve what supports it, and do not let a system flatten the thing that matters.
05 / OPERATING BACKGROUND
More than two decades in consequence-bearing operations.
Before building AI and decision software, I spent more than two decades in multifamily real estate, asset management, portfolio improvement, training, turnarounds and high-volume operations. That work spanned different markets, ownership structures and operating environments and shaped the way I read variance, sequence, physical evidence and execution risk.
OPERATING EXPERIENCE
Portfolio + asset operations
Multifamily and asset-management work across acquisition assumptions, repositioning, stabilization, receivership/disposition, operating strategy and portfolio improvement.
OPERATING EXPERIENCE
Value-add + turnaround execution
Hands-on experience with distressed and underperforming assets where plan-versus-actual variance, physical condition, capital sequencing and operating follow-through directly affected outcomes.
OPERATING EXPERIENCE
Scale + operating systems
Built and supported operating structures for high-volume portfolios, including leasing, training, field execution, reporting and cross-functional coordination.
OPERATING EXPERIENCE
Decision-making under incomplete information
Much of the work required acting before every answer was available: separating what was known from what was assumed, identifying the missing source, and recognizing when several small signals changed the route.
OPERATING EXPERIENCE
Pattern recognition across markets
The operating environments changed, but the recurring work did not: notice variance early, preserve context, understand sequence and avoid flattening a complicated asset or operating problem into one number.
OPERATING EXPERIENCE
From operations to software
Kingan Logic grew from trying to make those same reasoning habits explicit in software—especially where evidence, state, authority and next action need to remain separate.
06 / SELECTIVE ACCESS
External relationships should buy down uncertainty.
I am open to selectively scoped paid pilots, data-bearing evaluations, research replication, infrastructure partnerships and aligned strategic conversations when the outside relationship creates a real proof class, operating environment or commercial learning opportunity. Pilot design is specific to the organization and use case rather than one standardized package.
Research identity / contact
Independent research.
Real operating problems.
Current research interests include AI evaluation, human–AI decision systems, decision integrity, benchmark validity, state preservation, correction burden, operational intelligence and domain-specific predictive systems.
Public scholarly outputs and research profiles are maintained on the Research page.
Contactlogan@kinganlogic.comLinkedInResearch
Kingan Logic / Max R.O.I. LLC
Birmingham, Alabama, United States