Strategy. Software. Systems.
Intelligent systems. Built around your business.
Willowark engineers across the boundary between digital and physical systems — AI automation, custom software, machine vision, IoT, and industrial integration.
Why we exist
Most problems live at the boundary.
A machine that isn't connected to anything.
It runs all day. Nobody knows how well until the end of the shift — and by then the scrap is already in the bin.
A sensor is cheap. Knowing which one is the job.
A current transformer on the drive, a proximity switch on the ram. Choosing the right one, and where to put it, is engineering.
Something has to speak both languages.
The sensor is analog, the PLC talks Modbus, the office wants JSON. An edge gateway sits in the middle and translates — and buffers when the network drops.
Data that nobody can act on is just storage.
A broker and a time-series store turn readings into something that can be queried, trended, and compared shift to shift.
The answer has to reach a person who can act.
A dashboard on the plant TV. A phone that buzzes when the count falls behind. Now the shift can change while it still matters.
One partner across all of it.
Software firms won't touch the plant floor; controls integrators don't build cloud platforms; AI consultancies have never wired a sensor. Willowark works across sensing, software, data, AI and infrastructure — one accountable engineering partner instead of four vendors pointing at each other.
Hover or focus a component to see what it is and what it talks to. Arrow keys move between them.
A machine that is not connected to anything becomes a monitored, alerting system: sensor, gateway, broker and store, dashboard, and a phone that rings when something is wrong.
Components:
- The machine (not connected): Runs all day. Nobody knows how well until the end of the shift.
- A sensor (the right one): A current transformer on the drive, or a proximity switch on the ram. Cheap. Choosing it is the job.
- Edge gateway (speaks both): Reads the sensor and the PLC, speaks MQTT upward, buffers when the network drops.
- Broker + store (every reading): Pub/sub in, time-series out. Data nobody can act on is just storage — this is where it becomes something else.
- Dashboard (the shift, live): Counts, downtime, reasons — on the plant TV and in the office.
- A phone (someone who can act): The answer reaches a person while it still matters.
Connections:
- A sensor to Edge gateway over 4-20mA
- The machine to Edge gateway over Modbus
- Edge gateway to Broker + store over MQTT
- Broker + store to Dashboard over REST
- Broker + store to A phone over push
A typical monitoring system, assembled as you read. Hover or focus a component; arrow keys move between them.
What we build
Eight capabilities. One engineering partner.
- 01AI & Intelligent AutomationAI agents, workflow automation, LLM integration, and knowledge systems that take repetitive work off your team's plate.Learn more →
- 02Software EngineeringCustom software from rapid prototype to production — full stack, backend, APIs, data systems, and internal tools.Learn more →
- 03Cloud & InfrastructureCloud architecture, DevOps, deployment, and edge-to-cloud infrastructure that keeps systems running.Learn more →
- 04Industrial AutomationManufacturing automation, machine integration, IIoT, production monitoring, and industrial AI for the plant floor.Learn more →
- 05Vision & Advanced SensingMachine vision, AI inspection, laser measurement, 3D vision, radar, X-ray integration, and sensor fusion.Learn more →
- 06IoT & Smart SystemsRaspberry Pi and embedded systems, smart sensors, electrical and energy monitoring, gateways, and edge computing.Learn more →
- 07Systems IntegrationWe make technologies that weren't designed together work as one system — hardware, software, machines, and cloud.Learn more →
- 08R&D & PrototypingProofs of concept, product engineering, test systems, and experimental builds — from technical idea to working prototype.Learn more →
Selected work
Engineering that holds up in production.
Radar sensors read strip position, an edge controller turns raw returns into an offset, and the mill PLC drives the centering actuators while the operator watches live position.
A steel-mill systems provider
Radar-based centering system for steel mills
Steel manufacturing
An X-ray source and detector feed an acquisition chain; gauge software applies a calibrated thickness model and passes thickness back to line control.
An industrial measurement company
X-ray thickness gauge
Industrial measurement
Schedulers work in operating software with AI assistance, backed by distributed cloud services and connected to a training game.
A global automotive manufacturer's training operation
AI operating software for training operations
Automotive
Legacy points-and-rewards systems migrate onto event streams built as Go microservices, feeding a ledger of balances and global member APIs.
A global loyalty & rewards platform
Global loyalty & rewards platform modernization
Financial services
Work it out yourself
What is unplanned downtime costing you?
Your units, your margin, your idle labour. Computes here, in the page; nothing is sent. Five more instruments — inspection payback, vision feasibility, legacy connectivity, automation payback, and a sketch of your own system — are one click away.
Good output when the line runs.
Contribution margin, not sale price.
Stops nobody planned: faults, jams, waiting on parts or people.
Everyone paid while the line is down, loaded rate.
Overtime, rush freight, scrapped work-in-progress.
What monitoring plus the top two fixes typically remove. Your number, not ours.
Cost per down hour
Lost margin plus idle labour.
Per month
Per year
216 unplanned hours a year.
A 20% reduction is worth
Per year. This is the number a monitoring project has to beat.
How this is calculated
Cost per hour = units/h × margin + idle labour.
Monthly = unplanned hours × cost per hour + expedites and scrap.
Annual = monthly × 12. Reduction value = annual × reduction %.
Planning numbers, not a quote — a written scope follows a conversation.
Straight answers
Describe the problem. Get a straight answer.
One line is enough — what is manual, what is not connected, what you are trying to build. An engineer reads it and replies within a business day with whether and how we would approach it.
How an engagement works
Straightforward, start to finish.
01
Tell us the problem
A call or a message — describe what's manual, disconnected, or missing. An engineer reads it, not a sales script.
02
We scope it
On-site or from your systems and drawings — we work out what the solution actually takes, and say so if it doesn't need us.
03
Clear proposal
A written scope, deliverables, and pricing model before work starts. No moving targets.
04
Build & integrate
Software, hardware, and integration engineered together — with working checkpoints along the way, not a black box.
05
Commission & support
Deployed into your real environment, verified, documented, and supported after handover.
Who we work with
Built for the people doing real work.
Small & Medium Businesses
We automate expensive manual business processes.
How we help →Manufacturers
We make your factory smarter.
How we help →OEMs & Machine Builders
The software, AI, vision, IoT and systems engineering behind your machines.
How we help →System Integrators
Specialized engineering capacity without adding headcount.
How we help →Hardware Startups
One engineering partner from prototype to production.
How we help →Across 33 industries, from steel mills to plumbing companies — the same expensive problems, different vocabulary →
Where we work
Rooted in the Capital Region. Working nationwide.
Based in the Saratoga Springs area of New York, on-site across the Capital Region — from the semiconductor corridor in Malta to the shops and plants of Albany, Troy, Schenectady, and Glens Falls. Everywhere else, engineering runs remote with travel for commissioning and installation.
Explore service areasOn site in 33 Capital Region communities — Saratoga Springs, Malta, Clifton Park, Albany, Schenectady, Troy, Glens Falls and the rest — and remote-first in every state. Find yours →
Plain answers
Questions, answered plainly.
What does Willowark actually do?
We engineer systems across the boundary between digital and physical operations: AI automation for business processes, custom software, machine vision and inspection, industrial automation and machine data, IoT and edge systems, and the integration work that makes separately-designed technologies behave as one system.
Who does Willowark work with?
Small and mid-sized manufacturers, industrial OEMs and machine builders, system integrators who need specialized engineering capacity, hardware startups, and operations-heavy businesses drowning in manual processes.
Where does Willowark work?
We're based in the Saratoga Springs area of New York and work on-site across the Capital Region. Nationwide, software, AI, and cloud engineering runs remotely, with on-site travel for commissioning, integration, and installation.
How do engagements start?
With a conversation and a written scope. For larger efforts we often propose a small first phase — an assessment or proof of concept — so both sides learn what the full build takes before committing to it.
Insights
From the blog.
Not ready to talk?
Start with a working guide.
Curated engineering notes and a checklist you can take into your next planning meeting.
Book a callFree guide
The Plant Data Starter Guide
See what your machines actually did today — without replacing any equipment.
Get the guide →