Work orders, not spreadsheets
Work-order, routing and station workflows with role-based operator interfaces.
MES / ERP / Software that runs the floor
Software that runs the factory floor.
I built and deployed the MES and ERP running Western Magnetics' production, quality and supply workflows, working with production staff from first prototype through rollout.
Austin, Texas · Open to manufacturing roles
Matthew RundleBuilder. Team leader. Hands-on.
ManufacturingMES + ERP
Built from scratch and deployed at Western Magnetics.
End to end
Production, quality, purchasing, inventory and shipping in one system.
This résumé
Manufacturing Software and Applied AI
Western Magnetics, 2026 to present
How I work
Built around how the work happens.
Equipment in the loop
Interfaces that carry live machine signals into the MES, alongside the mechatronics engineers who built the controllers.
Traceable by default
An event stream records production state changes, from incoming inspection to shipment.
From the floor to the business
- FLOORFactory devicesMachines, controllers and test stations
- EDGEEdge deviceCarries machine signals into the MES
- MESExecutionWork orders, routings, quality, lots
- ERPBusinessPurchasing, inventory, shipping
Selected work / 01—03
Built to be used.
Three systems, three different businesses. Ownership from the problem through production.
The software that runs Western Magnetics’ factory.
I built the manufacturing execution system (MES) and ERP the company runs on: every build, quality check, purchase, inventory move and shipment goes through it.
The setting
A hard-tech startup that needed software built around the factory’s real operating needs, with the people doing the work.
Next.js · TypeScript · PostgreSQL · Prisma · Docker
01
Quality checks, nonconformance records and per-unit testing linked to the build.
02
Material tracking from incoming inspection through lots, FIFO, BOM consumption and kitting.
03
An agentic failure-analysis workbench connecting test results to evidence-linked conclusions.
INSIDE THE SYSTEM
A working view of manufacturing.
Explore the modules, select a record, and follow the information behind a build.
MANUFACTURING EXECUTION
Work orders
Follow each build from released materials to a tested motor.
| Work order / product | Operation | Status |
|---|---|---|
| 020 / Winding | In progress | |
| 030 / Assembly | Queued | |
| 040 / Balance & test | Quality hold |
WO-1042 / Record detailTRACEABLE BY DESIGN
Material lots, equipment and routing revision stay attached to the production record.
- Operations definition
- MOTOR-28 / Rev C
- Work center
- Motor cell 01
- Equipment
- Winder W-01
- Material lot
- LOT-CU-019
- Planned / completed
- 120 / 84 ea
- Personnel class
- Winding-qualified operator
Portfolio concept using fictional orders, lots and quantities. Illustrates the workflow without exposing a production system.
FROM THE FLOOR TO THE BUSINESS
Factory devices. Edge integration.
Cloud-connected operations.
Factory devices connect through an edge device to the MES. Tablet interfaces are part of the MES, which connects to the cloud-hosted ERP.
FACTORY DEVICESMachines · controllers · sensors
01WindingEquipment signals & results
02Press / assemblyEquipment signals & results
03BalancingEquipment signals & results
04Electrical testEquipment signals & results
Equipment connections Machine states, measurements & test results
ON SITE / EDGE
Edge device
The connection point between factory equipment and the MES
Equipment data into the MES MQTT as a messaging design principle
MANUFACTURING OPERATIONS
MES
Work orders · routings · material genealogy · quality · production records
Tablets / operator interfaces
Part of the MES: where operators interact with the manufacturing workflow.
MES ↔ ERP Versioned API contracts & defined data ownership
CLOUD-HOSTED
BUSINESS OPERATIONS
ERP
Orders · purchasing · planning · financial inventory
ISA-95
A shared model for operations, materials, equipment and personnel. Routings and BOMs connect to versioned definitions and resource requirements.
Enterprise / control integration ↗Purdue model
Use the separation of business operations, manufacturing operations and equipment control to reason about responsibilities. The diagram shows the device-to-MES-to-ERP path, rather than a literal Purdue network stack.
Industrial network layers ↗MQTT
Publish equipment events through a broker so consumers can subscribe without coupling directly to each machine.
Publish / subscribe messaging ↗API-first contracts
Define schemas, versions and error behavior at service boundaries. Producers and consumers can evolve against an explicit contract.
Amazon / AWS service-contract guidance ↗Simplified view: factory devices → edge → MES → cloud-hosted ERP. Tablet operator interfaces belong to the MES. Design principles describe the approach; they do not specify every deployed service or network control.
AI that handles a small business’s calls, texts and email.
One platform for outreach and follow-up. AI agents make and answer calls, work the inbox and book appointments, then turn conversations into orders.
- Verifies conversation outcomes before creating follow-up tasks.
- Detects stalled work, bounds retries and escalates to a person when needed.
- Drip campaigns, cold and warm-lead calling, AI inbox handling and scheduling through one agentic process.
Next.js · TypeScript · Supabase · Inngest · Retell · Twilio
AI that writes the dentist’s chart notes.
After a patient visit, Scribe drafts the clinical note and billing codes for the dentist to review and approve. It also tells the practice what supplies upcoming appointments need.
- AI-drafted SOAP notes, ICD and CDT codes and procedural notes.
- Clinicians edit, review and approve the drafts; review is built into the workflow.
- Connects upcoming appointments to inventory and practice operations.
Next.js · React · Supabase · Anthropic · OpenAI
Experience
Build it. Lead it.
Understand the business.
Expedia Group / Vrbo · 2019–2025
Data is useful when it changes a decision.
I built and managed a team of six data scientists, working with executives on the commercial decisions behind a travel business.
- Build the team. Shared Python and SQL standards for analytics work.
- Understand the market. Segmentation across more than 15 variables to prioritize high-value property acquisition.
- Inform the executives. Integrated disparate data into executive reporting for strategy and resource allocation.
2026 — Present
Western Magnetics
Manufacturing systems and applied AI
Built and deployed the MES and ERP running production, quality, purchasing, inventory and shipping.
2025
Domain Labs
Founder, AI and SaaS Solutions
Designed and built custom SaaS for small businesses, including CoreLinq Communications and CoreLinq Scribe.
2021 — 2025
Expedia Group / Vrbo
Senior Manager, Supply and Commercial Data Science
Built and managed a team of six data scientists. Established executive analytics for commercial strategy and resource allocation.
2019 — 2021
Expedia Group / Vrbo
Manager, Analytics and Data Science
Market segmentation across more than 15 variables; regional reporting for sales visibility and prioritization.
2013 — 2019
HomeAway · National Instruments · Advisory Board
Analytics and consulting
Global reporting, data analytics consulting and business analysis.
Toolkit
Manufacturing
MES and ERP, work orders and routings, equipment integration, quality records, material tracking and traceability
Engineering
TypeScript, Python, SQL, Next.js, React, PostgreSQL, Prisma, Docker and API integration
Applied AI
LLM APIs, agent workflows, evidence-linked agent traces, review and hardening of AI-generated code
Fair questions
What people
usually ask.
Did you build the equipment controllers?
No. Mechatronics engineers built the controllers. I built the interfaces that carry their live machine signals into the MES, and the workflows that use them.
How did the system get adopted on the floor?
I worked directly with production staff from the first prototype through integration, hardening and rollout, translating how they actually worked into the software they use.
What does traceability look like?
Work orders, routings and stations write to an event stream that records production state changes. Quality records, per-unit tests and material lots link back to the build.
Where does AI fit?
Two places: AI coding agents during development, with me owning review and hardening, and an agentic failure-analysis workbench that ties conclusions back to source evidence.