From Terraform to the pixel
One team carries it end to end — infrastructure, backend, AI and interface by the same hands. Which means nothing falls between contractors, and the person who built the API is the person who knows why the screen is slow.
105+
endpoints behind 13 routers on the platform we are building
5
roles in the access model, enforced at the data layer
0
cross-tenant leakage, proven by a script rather than promised
Four builds, described the way the record allows
A full multi-tenant platform in active build
A modern React and TypeScript front end — 31 pages, 69 components, knowledge-graph visualisation, document preview, light and dark, mobile-ready — over a FastAPI backend with 13 routers, 105 endpoints and streaming responses, behind five-role access control and zero-trust tenant isolation.
Architecture leadership across three platforms
For a SaaS group running auction, appraisal and dealer-website products: one unified domain model on a shared multi-tenant core, API governance, and technical direction of several developer teams.
Product surfaces that shipped
A Word add-in lawyers work in daily plus an operations dashboard, both part of our own contract-review product.
Product sites, end to end
Including this one and our product site — front end, content model, and the whole machine layer: the schema graph, llms.txt and raw twins.
Multi-tenancy proven, not promised
A forgotten scope should return zero rows, never somebody else's data.
That is a property you can test rather than a promise you have to accept: our Quote Rabbit demo ships a script that proves row-level isolation by trying to breach it, and the same pattern goes into what we build for you. When an enterprise customer's security review asks how isolation is enforced, the answer should be a policy in the database and a test that fails loudly.
What we build into a product
Multi-tenant AI platform
Per-org tool registry, isolation at the data layer, credential vaulting and metered AI spend.
See it →AI agents
The layer that does the work inside the product, with guardrails written in code.
See it →Cloud and DevOps
The environments, pipelines and rollback path underneath, defined as code you keep.
See it →Bring the product and the date
This is our AI engineering practice
It is real work and it is where our four products came from. But what Cognilium leads with is narrower: optimization apps that run in tandem with Microsoft Dynamics 365, computing the decisions the ERP records but does not derive — the optimal price, the optimal pick path, the optimal stock level. See the optimization apps · How we build inside the ERP.