Services

We build the hard part and teach you to run it

Four ways to work with us, one way of working: make the problem measurable, build the simple thing that survives production, and make sure your team can carry on without us.

What we do04

Four ways to work with us

We usually start in energy, but the craft is the same everywhere: make the problem measurable, build the simple thing that survives, and hand over the knowledge.

Energy optimization

We build control that acts on price, forecast and measurements instead of rules of thumb. Plan the day, respect the power ceiling, and recompute when reality drifts from the plan.

  • Linear optimization
  • Forecasting
  • Peak shaving
  • Modbus & OCPP

AI that survives production

Models are easy to demo and hard to run. We build them with measurement, limits and fallbacks from the start — so they can be trusted on a Tuesday in November.

  • Forecast models
  • Anomaly detection
  • LLM agents
  • Evaluation

AI mentorship

We sit next to your team and build the first solution together. Your code, your repo, your decisions. We hand over the capability, not just a delivery.

  • Pair programming
  • Architecture
  • Code review
  • Workshops

Software engineering

The platform under everything else: cloud, APIs, real-time data, apps and hardware integrations. The same craft we have delivered as consultants since 2018.

  • Cloud & APIs
  • Real-time
  • Apps
  • Integrations
Where we work

Energy first — but the craft travels

We spent the years since 2018 building systems for other people. Now we have a field of our own where we know more than most, and we bring the same way of working into other industries.

Home ground

Energy

Energy management, battery optimization, forecasting, metering data and integrations with inverters, meters and chargers.

Adjacent

Industry and property

Peak-power control, operations monitoring and decision support where the measurements already exist but nobody acts on them.

Broader

Product and platform

Consulting in other industries: cloud, APIs, apps, and the AI capability you want inside the building.

AI mentorship

We're done when you no longer need us

Most teams don't need an AI vendor. They need someone to build the first solution with them and then step aside.

  1. Map it

    We go through your data, systems and ambition, and give an honest picture of what AI solves for you — and what it doesn't. You get a prioritised list, not a pitch.

  2. Build together

    We build the first solution alongside your team, in your repo. Pair programming, architecture and review happen in the work rather than in a report.

  3. Hand over

    We step down. What remains is documentation, tests and a team that builds the next solution on its own. We stay available as a sounding board for as long as it helps.

How we work

Three ways to bring us in

ModelFits whenShape
Build with usYou have a team and want to move faster. We build side by side and hand over the capability as we go.Ongoing · part-time
Build for usYou have a clear problem but no capacity. We take the whole delivery, and the handover with it.Project · fixed scope
Sounding boardYou have already started and need someone to review the architecture and hold the direction.Advisory · hourly

Tell us what you're trying to solve

One conversation is enough to tell whether we're the right fit. If you already have data and a system that should get smarter, that's usually where we start.