MNEOS — Institute for Memory & Intelligent Systems
MNEOS is a computational engineering institution.
We connect scientific knowledge, AI, simulation, experiments, manufacturing, and institutional memory so difficult engineering problems can be solved faster — and the knowledge gained is not lost.
Current build status
Updated July 2026
DOS memory & evidence substrate
The MNEOS memory and provenance layer, in daily internal use.
Federal engineering programs
Multi-program execution across USMC, NAVSEA, and ONR customers; details on request.
Research
Computational engineering methods, ceramic additive manufacturing, RF materials, memory architecture, and mathematical foundations.
Public learning resources
The Learning hub, Why MNEOS lineage, Computational Engineering, and the Library (memory, intelligence, institutions, and adjacent subjects) are live and continuously extended.
In developmentDedicated research pages on the lineage of inquiry, foundations of reasoning, and mathematical sciences.
PlannedInteractive engineering demonstration — an end-to-end worked example showing how MNEOS reasons from question to consequence.
The long-term architecture is ambitious. The development sequence is disciplined. MNEOS distinguishes between institutional principles, current capabilities, active development, and long-term architecture, and does not blur those categories. What is described here is what is operating today or being implemented through active work; larger capabilities appear on this page as they land.
Built on
Three decades of advanced engineering and manufacturing.
MNEOS founder David W. Sherrer founded Haleos (acquired 2002) and Nuvotronics (acquired 2019), led DARPA Phase I–III execution on the PolyStrata® RF microsystems platform, built qualified manufacturing capability, and is an inventor on 125+ issued patents worldwide.
How MNEOS works
Data flows to memory. Memory shapes understanding. Understanding produces coordinated action.
This is the substrate MNEOS is building. Every institution already generates enormous volumes of information. Very few can convert that stream into governed institutional intelligence that reaches action.
- DataSignals from people, instruments, systems, and environments.
- MemoryGoverned institutional recall with evidence and provenance.
- UnderstandingPatterns and causal structure across disciplines.
- InsightRecommendations, alternatives, and decisions surfaced to the institution.
- ActionCoordinated execution routed to the right people, systems, and instruments.
- Question — someone with judgment decides what to ask.
- Knowledge and evidence — what the institution already knows is surfaced first.
- AI proposal — candidate approaches, attributed and reviewable.
- Human judgment — a person evaluates the proposal and authorises what happens.
- Simulation and test — the proposal meets physics and measurement.
- Captured memory — question, reasoning, decision, evidence, and outcome are preserved together, and become the starting point of the next question.
AI proposes. Humans judge. Evidence constrains.
Every action moves through this doctrine. AI can suggest; only humans decide; every claim carries its evidence.
Memory precedes intelligence.
Structured, governed memory is the substrate that makes learning possible. Nothing important disappears when a person leaves.
Disciplines converge.
Physics, materials, manufacturing, sensing, and AI operate against a shared substrate rather than through document handoffs.
MNEOS in practice
What the system does, concretely.
These are examples of the kinds of problems MNEOS is designed to work on. Some are already underway. Others are what the platform will make possible as it matures.
Advanced Materials
Foam that behaves like a designed material, not a chemistry accident.
Physics-first modeling, instrumented pilot lines, and captured process reasoning let a new material class move from concept to reproducible manufacture without losing the engineering intent along the way. Currently underway on the TacFOAM program.
Live · TacFOAM Phase II
Human-AI Engineering
Voice-to-CAD and design intent that survives the conversation.
An engineer describes what they need. AI proposes geometry and material choices. Simulation checks the physics. The reasoning is captured with the artifact. When someone else picks it up months later, the ‘why’ is still there.
In active development
Biological Experimentation
Experiments that build on each other instead of restarting.
Wet-lab results, models, and interpretations enter one memory. The next experiment can reason over every prior result. The institution stops re-discovering things it already knew.
In active development · ONR Sleep as reference case
Defense-Technology Transition
Allied innovation carried into U.S. capability without losing the originator.
MNEOS provides the memory and evidence substrate that lets Ukrainian and other allied technologies move through the transition pathway with attribution intact, technical reasoning preserved, and governance visible to program offices.
Operating via Helicon Defense
Explore
Explore MNEOS.
Four ways in, depending on what you came for.
Participate
Work with MNEOS.
We work with a small number of collaborators at a time. If any of the following fits you, we would like to hear from you.
Researchers
Collaborations on computational engineering, memory systems, RF materials, or ceramic additive manufacturing.
Universities
Joint research, student projects, and access to the MNEOS platform for teaching and learning.
Industry partners
Technology transition, contract research and development, licensing, and joint program pursuits.
Government sponsors
SBIR and STTR execution, program transition, and mission-focused engineering.
Students
Internship inquiries, thesis collaboration, and mentored open-problem work.
Engineers
Hiring opportunities at MacroVation and affiliated ventures.
Potential partners
Anyone else who reads what MNEOS does and wants to talk about how we might work together.