UNDERSTAND
Know what exists
Turn geometry and engineering data into structured, traceable physical knowledge.
STRATIC SYSTEMS RESEARCH
Teaching AI to understand physical reality.
Our research asks how measurable engineering evidence can become useful intelligence for both AI systems and the people working beside them.
Reality
grounded facts
Geometry
measured evidence
People
expanded capability
Systems
connected work
RESEARCH CORE
Engineering
Intelligence
REALITY
grounded
GEOMETRY
measured
PEOPLE
empowered
SYSTEMS
connected
PEOPLE + AI + ENGINEERING
OUR RESEARCH THESIS
Language can describe an engineered system. Understanding requires a model of what is actually there.
A model can explain what a beam, weld, hole, tolerance, or assembly is. Engineering intelligence must go further: identify the actual entities, locate them, measure them, represent their relationships, and distinguish observed evidence from inference.
The same knowledge should also help people work better. We want systems that can teach concepts in context, guide real engineering tasks, validate work, and reduce repetitive process steps without hiding why a conclusion was reached.
EMPOWER
Increase human capability
Teach, explain, validate, and automate inside the workflow while preserving human authority.
CONNECT
Move intelligence through the work
Link CAD, project, manufacturing, and business systems through deliberate integrations.
ACTIVE RESEARCH PROGRAMS
Four connected questions. One engineering-intelligence direction.
Our research moves from physical representation to human capability and the systems required to apply that intelligence in real engineering work.
01ACTIVE RESEARCH
Physical Reality
How can AI move from describing engineered objects to understanding what physically exists?
We study representations that give AI a grounded view of engineered objects, spatial relationships, assemblies, features, and measurable physical facts.
CURRENT FOCUS
Engineered-object identity
Spatial and assembly relationships
Physical-world grounding
02ACTIVE RESEARCH
Deterministic Geometry
How should geometry become structured, queryable knowledge instead of an opaque visual signal?
We develop deterministic methods for turning CAD and engineering geometry into entities, features, frames, measurements, and relationships with traceable evidence.
CURRENT FOCUS
Canonical coordinate systems
Feature and entity extraction
Traceable geometric evidence
03ACTIVE RESEARCH
Human-AI Engineering
How can AI increase human capability instead of simply trying to remove people from the workflow?
We explore systems that teach concepts in context, explain engineering evidence, guide decisions, validate work, and automate repetitive steps while keeping people responsible for consequential decisions.
CURRENT FOCUS
In-context engineering education
Human-confirmed workflows
Evidence-aware assistance
04ACTIVE RESEARCH
Connected Engineering Systems
How can engineering intelligence move reliably between the software systems where real work happens?
We study practical integration patterns across CAD, project systems, manufacturing data, and business workflows so useful intelligence is not trapped inside one application.
CURRENT FOCUS
Structured system integration
Traceable workflow state
Reliable engineering handoffs
INTERNAL RESEARCH PROGRAM
GeometryLab builds capabilities behind our engineering tools.
GeometryLab is Stratic Systems' internal, research-grade program for building and testing capabilities in CADSync and other systems. It studies engineered geometry, features, measurements, and relationships while preserving the evidence behind each result.
The goal is to give our engineering tools a more reliable basis for review and explanation. GeometryLab is not a public product; methods and recognition coverage are still being qualified.
STRUCTURED ENGINEERING EVIDENCE
geometry / relationship / source
- ENTITY
- engineered object
- FRAME
- canonical / local
- AXIS
- measured orientation
- FEATURES
- holes · slots · copes · notches
- RELATIONS
- adjacent · aligned · connected
- SOURCE
- evidence retained
MEASURE
explicit
RELATE
structured
EXPLAIN
traceable
RESEARCH PRINCIPLES
Intelligence should make both the system and the user more capable.
The objective is not automation for its own sake. It is engineering work that becomes more understandable, more teachable, more connected, and more reliable.
Deterministic first
Measure and structure what can be established directly before asking a model to infer what remains uncertain.
Evidence stays attached
Results should preserve source geometry, transforms, relationships, assumptions, and confidence so they can be inspected.
Teach while assisting
Engineering software should help users understand why a result matters, not merely automate the next click.
Human authority remains
Consequential engineering decisions require visible evidence, review, and clearly defined responsibility.
Connect deliberately
Integrations should move the minimum necessary information across well-defined system boundaries with traceable intent.
PUBLISHED RESEARCH
Research notes on engineering intelligence and physical reality.
These papers document the working research behind Stratic Systems: deterministic evidence beneath AI reasoning, explicit provenance, human authority, and useful engineering capability.
018 min read
PHYSICAL INTELLIGENCE
Physical Intelligence: Moving AI from Describing Objects to Understanding Them
A Stratic Systems research perspective on the representations AI needs to move from fluent descriptions of engineered objects to traceable understanding of physical systems.
Jeffery C. Wheat · Aug 6, 2026→
029 min read
DETERMINISTIC GEOMETRY
Why Industrial AI Needs a Deterministic Geometry Layer
Why industrial AI systems benefit from separating measurable geometric facts from probabilistic interpretation and language-model reasoning.
Jeffery C. Wheat · Aug 6, 2026→
039 min read
KNOWLEDGE PRESERVATION
Capturing Engineering Knowledge Before It Disappears
A practical framework for preserving engineering knowledge that lives in people, geometry, drawings, decisions, and manufacturing practice before organizations are forced to rediscover it.
Jeffery C. Wheat · Aug 6, 2026→
049 min read
CAD INTELLIGENCE
From CAD Files to Queryable Engineering Intelligence
A research architecture for turning CAD from a passive file opened by one application into a structured, traceable source of engineering knowledge that software and AI systems can query.
Jeffery C. Wheat · Aug 6, 2026→
059 min read
ENGINEERING REASONING
Trustworthy AI for Manufacturing Requires More Than a Language Model
Why industrial AI systems need evidence separation, explicit uncertainty, human authority, deterministic tools, security, and evaluation around the language model itself.
Jeffery C. Wheat · Aug 6, 2026→
068 min read
PHYSICAL SYSTEMS
Connecting CAD, Field Vision, and Physical Systems Through STX
A Stratic Systems research proposal for using STX as a portable engineering context package between CAD-derived physical knowledge and field vision systems.
Jeffery C. Wheat · Aug 6, 2026→
PUBLIC RESEARCH · PROTECTED ENGINEERING
Share the direction. Protect private engineering data and implementation.
Stratic Systems can publish research questions, high-level methods, selected findings, and demonstrations without exposing customer information or proprietary engineering assets.
PUBLIC
Research direction
High-level architecture
Selected validated results
Responsible demonstrations
PROTECTED
Customer and partner data
Private engineering assets
Proprietary datasets
Production implementation details
THE DIRECTION
Teach AI the physical world. Give people better tools to work in it.
That requires measurable geometry, engineering knowledge, human-centered assistance, and reliable connections between the systems where engineering work happens.