Engineering With AI

3 Disciplines.
Agentic workflows.
Built in public.

AI-augmented engineering methodology applied across electrical, mechanical and software disciplines.

Watch the build → Work together
Scroll
3
Disciplines
one methodology
12
Years
engineering
DoD/DoW
Active
clearance
17 OSS
Artifacts
Github
Methodology

Self-corrected engineering
is the differentiator.

The discipline is in the validation gates and composable primitives that catch failure across electrical, mechanical, and software disciplines.

By creating a predictable environment and giving the agent means to verify its own output at each gate, you get the speed of AI generation without inheriting its failure modes.

01

Composable Primitives

Infrastructure decisions decomposed into independently verifiable units — each with defined interfaces and failure modes across electrical, mechanical, and software disciplines.

02

Validation Gates

Every integration point has a gate. The model can't hack its way around a gate. Agent works iteratively until validation and then moves on to the next step.

03

Convention As Constraint

Models are trained on the patterns that appear most in human-written code — MVC, REST conventions, component architecture, standard module layouts. When your codebase follows those conventions, the agent isn't guessing how to extend it. Convention isn't just organization — it's the language the model already speaks.

From requirements
to fabrication/production

Electrical, mechanical, and software templates from the Engineering With AI methodology. Each one is a solved problem class — validation gates, simulation harness, and agent system prompt included. Each one showcasing a powerful agentic workflow.

View all templates on GitHub →
Sample templates — 3 of 19
⚡

[elec] computer vision sensor board

SKiDL netlist, PySpice simulation, KiCad layout.

→
🔩

[mech] custom bracket connector

CadQuery parametric CAD, CalculiX FEM, GD&T drawing export.

→
🤖

[software] python-domain-mcp-server

Software · Neo4j graph + pgvector RAG, fact verification, FastMCP tools

→
Live Project

Here's the methodology applied to a real project — all three disciplines converging on one system.

Current Project

ARCNODE

Modular AI infrastructure. Physical shipping containers in the following flavors: Energy Storage (BESS), AI compute, Thermal and Grid. Customer inputs deployment requirements into the System Configurator and receives a full engineering package: BOM, CAD, installation and manufacturing docs, plus an AI powered dashboard to run the system. Built with EWA templates.

Off-Grid AI Inference Defense / Sovereign BESS Thermal Power Systems IIoT
View ARCNODE →
ModulesBESS · Compute · Thermal · Grid Interface
Container10ft ISO Shipping Container per module
DeploymentOff-grid · Remote · Sovereign · Defense
ConfiguratorRequirements in → BOM + CAD + Docs + EMS out
Energy Management Dashboard (EMS)RAG/Graph analyst · Dynamic sensor-to-UI config
EDP APIEngineering Deployment Package: Doc generator for procurement and construction of datacenters and their energy needs.
Custom Interface PlatesCustom machined plates that provide cable and pipe connectors for arcnode module interface.
StatusActive — documented in public

Two ways in.

Same methodology. Different engagement — depending on whether you want to learn it or deploy it in production.

For engineers & technical teams

Learn the
methodology.

The ARCNODE project is documented on YouTube. Not a survey of AI tools — a working system for mechanical, electrical and software engineering that fully leverages the power of AI tooling.

  • Composable primitive design across disciplines
  • Validation gate implementation
  • AI-augmented iteration workflows
  • Real project — no manufactured examples
Watch on YouTube →
For operators, developers & defense

Deploy it
in production.

Implementation consulting for AI datacenter developers, microgrid operators, and defense/sovereign deployments. DoD clearance. Engagements scoped to what you actually need.

  • Off-grid AI inference infrastructure
  • Defense & sovereign resilient compute
  • Microgrid + BESS integration
  • AI-augmented engineering process adoption
  • Energy tech startup advisory
Inquiry only →
About
Joe Narvaez
Joe Narvaez

Software engineer by trade, energy engineer by degree. Twelve years building systems across electrical, mechanical, and software. I've always had a deep appreciation for the fundamentals/first principles. AI turned that into a force multiplier. Same depth, significantly greater range.

BS Renewable Energy Eng. Oregon Institute of Technology
MS CS — in progress Georgia Tech · ML + Cyber Physical Systems
Tier 3 Secret DoD Clearance
EIT — #155352 ABET California · 2015
Worldwide Onsite Contracts International long term forward deployments available
Consulting

If it sounds like a good fit,
let's chat.

If your problem fits the methodology — off-grid AI compute, defense/sovereign resilience, energy tech architecture — send a brief description and I'll tell you whether I can help.

Defense and sovereign inquiries: include classification level in your message. Responses within 48 hours.