Skip to content

Engineering

I build the path data takes from the machine it comes off to the decision someone makes with it.

01

Sense

The data exists, but it is trapped in the machine.

What I build

  • Sensor selection and placement for the measurement you actually need
  • Signal conditioning and acquisition on industrial hardware
  • Calibration against a known reference, so the numbers mean something
A real signal is continuous and noisy long before it becomes a row in a table.

Control & instrumentation

02

Move

It is on the machine, in a vendor portal, and in a PDF emailed every Monday.

What I build

  • Telemetry paths off the plant floor
  • Vendor export ingestion, scheduled and monitored rather than remembered
  • LLM-assisted extraction from the documents nobody wants to retype
Sampling is a decision, and it throws information away. The trace stays on the machine; only the samples travel.

Industrial data acquisition — edge to cloud

03

Shape

Twelve sources, twelve schemas, no shared key, and nobody trusts the total.

What I build

  • Normalisation into one queryable model
  • Validation and reconciliation, with the mismatches surfaced instead of dropped
  • Provenance you can audit back to the source row
Normalisation is the act of making things comparable. The rows that refuse to line up are the job, not an edge case.

Data quality, sensor fusion

04

Decide

Someone still opens four dashboards and makes the call from memory.

What I build

  • Scheduled reporting that lands before the meeting, not after
  • Thresholds and alerting on the numbers that actually move
  • Monitoring surfaces and digital-twin views of the running system
All of it exists to produce one number a person acts on, and the moment worth building for is when it crosses a line.

Digital twins, Industry 4.0, closed-loop control

Then the decision goes back to the machine

It changes what the instrument reads, which changes the next decision. Studying that loop is the PhD.

Read the research

Stack

Sense
  • Sensors & DAQ
  • LabVIEW
  • Keil uVision
  • Altium
Move
  • Python
  • C++
  • REST APIs
  • LLM orchestration
Shape
  • Pandas
  • NumPy
  • SQL
  • PostgreSQL
Decide
  • MATLAB & Simulink
  • ETAP
  • Matplotlib
  • Scheduled jobs

Case studies

Working with me

Engagements start at any stage. Most start at Shape, because that is where the pain is loudest.

A first piece of work is usually two to four weeks, scoped to one measurable outcome we agree on before it starts.