The architecture was mine. The keyboard was someone else’s.

Solarex has been installing solar across Mauritius since 2013, with more than 20 MW active on the island. The engineering is their trade. The software around it, the public site and the back-office the team works in every day, needed rebuilding. What they did not need was an outside firm to take ownership of it. They had someone who could implement. What they wanted was someone to decide what to implement, and to be right about it.

Design the system, then hand it over. Every decision that is expensive to reverse was mine to make: the data model, how a lead travels from the public site into the business, what the people who live in the back-office all day actually need on screen. Then write it all down clearly enough that somebody else could build it without me in the room.

I made the decisions and I did not write the application.

The data model first, because it is the one thing you cannot quietly change later. Then the lead-capture engine: a visitor sizes their own system on the public calculator, and that becomes a qualified lead sitting in the back-office before they have closed the tab. Then the back-office itself, where the design question is never how it looks in a screenshot, but whether someone using it for the fortieth time that day still moves quickly.

Around the build, the parts that decide whether a team ships or stalls. Coding and architecture guidelines, so the implementer had a road instead of a blank page. CI/CD and deployment, set up by me so that shipping was routine from the first day rather than a ceremony. Search infrastructure, which is now bringing the business leads it was not getting before. Guidance on their email transition. And a standing seat on the calls, for the moments when the person building it needed a decision made rather than a suggestion offered.

The system is in production. The next phase, stock and inventory, is being specified now on the same terms.

A plan is only worth what somebody else can build from it. That is the whole test, and it is not the test most technical advice is written to pass. Solarex’s implementer built this one from a document, months after I wrote it, with me nowhere near the keyboard. What shipped is what was specified.

Solarex got a system in production that they own outright, built by their own implementer, with the expensive decisions already made correctly. And they came back for the next phase, which is the only review of advisory work that counts.

“Harivansh is straightforward to work with. He understood what we needed quickly, made clear decisions rather than presenting us with options, and stayed available whenever our developer had a question. We would work with him again, and we are.”

REFERENCE CALL AVAILABLE ON REQUEST