Software that does the boring parts well—so the rest can feel effortless.
We're six engineers and designers who build AI products and SaaS for companies who measure twice. Production-grade evals, edge deploys, and a senior team that stays through launch.
2023 — 2026
Four disciplines.
One small team.
We don't have a “growth” practice. We don't have a “discovery” practice. Just four things, done in-house, by the people you'll meet on day one.
AI products
LLM-native tools, retrieval, agentic workflows. Evals, observability, and cost controls from day one.
SaaS development
End-to-end product work. Auth, billing, ops — the plumbing that ships and the surface that delights.
Platform & infrastructure
Edge functions, multi-region data, queues, background jobs. The systems that quietly hold the product up.
Design systems
Component libraries and tokens that age well — built alongside the product, not sold separately.
Numbers we'd put on a contract, not a slide.
We track outcomes, not hours. Every engagement leaves with a public scorecard — what we promised, what we shipped, and the metric that moved.
“We've worked with three studios before Inventiva. This is the first one where the engineers cared about the same things we cared about.”
A four-beat rhythm. No waterfalls, no sprints theatre.
Six people. One office. No pipeline of juniors waiting to touch your product.
Inventiva was founded in 2026 in Lisbon by Sofia and Rafael after a decade of building at larger studios. We take on four or five engagements a year — no more — because we still want to do the work ourselves.
Notes from the studio.
We grow slowly, on purpose.
4 roles open right now. If none fit, leave a note anyway — we read every application.