Build analytics a regulator can check.
We build answers grounded in a model of India's grid, with the evidence for every figure, and a system that says so when the data cannot support an answer.
The work
One model of the grid
Meters, feeders, transformers, tariff orders and ratings in one ontology that every product reads from.
Evidence on every answer
Each answer carries its formula, inputs and source, and can be replayed to the same result.
Knowing when not to answer
The system must decline when the evidence is missing. That is harder to build than an answer.
How we work
On site
Engineers travel to the DISCOM and work with the people who use the output.
Decisions in writing
We record the reasoning behind every significant design decision.
Open-source stack
The deployed stack uses permissively licensed open-source software.
Honest labels
Every figure on screen is labelled live, modelled or illustrative, including in sales demos.
Open roles
Founding engineer: ontology and platform
The data model, the evidence record behind every answer, and the graph.
ApplyFounding engineer: grid ML
Theft detection, load flow and forecasting, with models that show their working.
ApplyForward-deployed engineer
On site at the DISCOM, running a pilot to a clear result.
ApplyProduct designer: decision screens
Worklists, evidence views and briefs for utility teams.
ApplyNo role that fits? Write to us anyway.