Summary
TRIS is a geospatial tool aimed at making carbon-credit origination accessible to Brazilian farmers and landowners — the segment of the market historically priced out of the certification process.
Brazil is one of the largest origination countries in the global voluntary carbon market, with credits coming from Amazon forest conservation, soil-carbon programs, and restoration projects. In late 2024 the country authorized a regulated domestic market (the SBCE) and a national certifier, both explicitly intended to lower certification cost and broaden access for smallholders. Most of the friction for small producers isn't the carbon — it's the geospatial baselining, MRV (monitoring, reporting, verification), and paperwork around it.
TRIS sits on that friction. The product packages the geospatial analysis a small landowner would otherwise have to commission from a consultancy, so the cost of finding out whether your land qualifies stops being a barrier.
What I worked on
- Built an early Unity 3D prototype to render reforestation projections in-browser via WebGL. It was abandoned during market-fit analysis, but it taught us what the product actually needed to be.
- Iterated from that prototype toward a console-based geospatial analysis pipeline, which became the core of the product.
- Wrapped the pipeline into a web UI tailored for farmers and landowners with no technical background.
- Built a separate data visualization dashboard for carbon consultants working across multiple properties.
- Designed a custom caching and querying layer to handle the massive geometries you get when working at Brazil's scale.
Tech and approach
The early prototype was Unity and WebGL. Once we moved past that, the stack settled into a GCP-based infrastructure with Google's geospatial services doing the heavy lifting on satellite imagery and land analysis. The backend was Python and Postgres for processing pipeline outputs and serving queries.
The front-end was a React SPA with Tailwind and Mapbox for map rendering. Most of the UI components were hand-rolled and iterated specifically for our audience. Farmers and consultants use software differently, and off-the-shelf component libraries didn't fit either group well enough.
Status
I moved on after we proved market fit and the company secured its initial funding from the state of São Paulo. The project is now in the hands of its CEO and founder, while I focus on work closer to where I live in Belgium.