Software engineering, AI, data and infrastructure. Built, modernized and made to work together, by an independent, founder-led engineering studio.
Meniva · Budapest · EuropeSoftware · AI · Data · Cloud · Modernization
01 · What we build
Every engagement is scoped to a deliverable you can point at. Most combine two or three of these.
All services →02 · Selected work
Project 01 · Client work
greehillvia
Greehill turns LiDAR scans of city streets into per-tree data for 130+ cities. For the production GPU-backed ML/data platform behind it, Meniva delivered an evidence-backed architecture, reliability and capacity assessment, then engineered the cross-repository preprocessing handoff. Delivered through WeSpeakAI.
Read the case study →Project 02 · In-house product
A full-stack learning application built and run by Meniva: structured lessons, executable practice, a learning map that tracks capability per concept, and an AI mentor embedded in the lesson flow. Architecture, implementation, instrumentation and iteration are all in-house.
Project 03 · Client work
For the Miutcánk civic initiative, Meniva analysed an online consultation on the redevelopment of Thurzó Park in Budapest. Thousands of residents answered in nine days; we turned their answers into a ranked, decision-ready picture of what the neighbourhood wants.
Read the case study →03 · How we work
Working software every week, decisions in writing, and code in your repositories from the first commit.
The full process →We read the code, the data and the constraints, and write down what the outcome has to be. Output: a scoped plan and estimate.
Architecture, data model and interfaces agreed before implementation. A thin prototype where the risk is technical.
Small, reviewed pull requests into your repository. Tests, CI and documentation are part of done.
Deployment, monitoring, runbooks and a handover your team can operate. Stay on a retainer, or not.
04 · Labs & experiments
Technical demonstrations with their boundaries stated. Not client work, not products: evidence of how we approach RAG, agents, scoring and local inference.
All experiments →05 · Contact
A few sentences about what you are building or what is in the way. You get a reply from the engineer who would do the work.
