Home/How we work
How we workFrom a real problem to software in production.
We don't sell an off-the-shelf product. We pick one operational problem in your energy or RES business, dig into the data and constraints behind it, and deliver working software fitted to how you actually run.
The engagement
Seven steps, fully transparent.
Every engagement runs the same way: understand the problem, agree the architecture, then ship and stay. No black box, no surprise scope — you see each stage and the data behind it.
We sit with the people who live the problem. We map the operational pain, the data sources behind it — SCADA, meters, ERP, market feeds — and the real constraints: grid codes, vendor lock-in, latency, downtime cost. The output is a sharp problem statement, not a wishlist.
We define a tight first scope and the architecture to carry it: data model, services, time-series storage, interfaces and where it plugs into your stack. Before a line of code is written, you know what ships first, what follows, and the protocols and systems involved.
We connect the sources — SCADA and PLC tags, smart meters and IoT devices, market and exchange data, weather APIs — and normalise them. Units, timestamps and vendor formats are reconciled into one clean, queryable feed. Most of the real work in energy software lives here.
We build the software against that feed: monitoring, forecasting, anomaly detection, control logic, reporting — whatever the problem needs. Short iterations, working increments you can see, and code written to run in production rather than demo well.
The software has to live inside your operation. We wire it into ERP and CRM, push KPIs into BI and reporting, and connect the data flow both ways so results land where your teams already work — not in another isolated dashboard nobody opens.
We test against real production data, not synthetic samples. Forecasts are checked against actual generation and load, anomaly rules against historical events, control logic in shadow mode before it acts. We tune until the numbers hold up on your sites, under your conditions.
We put it into production and stay. Observability, alerting and ownership are part of delivery, not an afterthought. As your portfolio grows and the grid changes, the software keeps pace — we extend it rather than hand you a frozen artefact.
Principles
Why it's built this way.
Event-driven & observable
Systems react to incoming data and state changes, and log every transition. When something drifts on a plant, you can see exactly where and when.
Built, not bolted on
We fit software to your market, processes and scale. No forcing your operation to match a generic product — the architecture follows your reality.
Security-aware
Energy is critical infrastructure. We treat access, network segmentation and OT/IT boundaries as design inputs, not a checklist bolted on at the end.
Delivery, not shelfware
We measure success by software running in production and earning its keep — not slide decks, pilots that never ship, or licences gathering dust.
Before the build
How messy energy data becomes a clean feed.
Meter, SCADA and market data arrive in mismatched units, timezones and vendor formats. Before any forecasting or control logic is worth running, the data has to be made trustworthy. This is the pipeline that gets it there.
Normalise
Units, timestamps and vendor formats are converted to one canonical schema — kW vs kWh, local time vs UTC, every device speaking the same language.
Deduplicate
Overlapping reads from redundant loggers, retries and re-imports collapse to one record per measurement, so totals and aggregates stay honest.
Validate & gap-fill
Out-of-range values and frozen sensors are flagged; communication gaps are detected and filled with documented, auditable methods — never silently.
Tag & map
Every series is mapped to its asset and site — string, inverter, plant, portfolio — so data can be rolled up, compared and filtered with confidence.
Aggregate
Raw high-frequency reads are resampled to 15-minute and hourly resolutions for settlement, reporting and forecasting, while raw data stays available.
Audit trail
Every transformation is recorded — what changed, when and why — so any number in a report can be traced back to the raw reading behind it.
Let's talk
Want to walk through your case?
Bring us one operational problem — a portfolio you can't see clearly, forecasts you can't trust, data trapped across systems. We'll tell you straight how we'd approach it.