A modeling practice that behaves like an engineering one.
Northern Virginia · Working globally
NexFlood exists to put deep, hands-on experience in coastal, riverine, and compound flooding directly against the hardest problems in the flood risk space — and to solve them with the best tools available rather than the ones that happen to be familiar. That means physics where physics is the right answer, machine learning where it earns its place, and a clear statement of which is doing the work.
The practice is run the way software gets built: under version control, with tests where tests are possible, and with everything needed to re-run and check a result shipped alongside the result. It costs about a week at the start of a project and pays for itself every time somebody asks how a number was produced.
We work directly with agencies, ports, and utilities, and as a specialist partner to flood technology companies that need modeling depth on a particular problem. Much of the work is international — South Asia, the Gulf, and other regions where exposure is high and instrumentation is thin.
Four shapes, and the honest length of each.
Scoping study
2–4 weeks
A written position on method, data availability, and defensibility, before the main budget is committed.
Modeling engagement
2–6 months
A full study: domain, forcing, simulation, validation, and a delivery package that opens without a walkthrough.
Systems build
3–12 months
A forecast or alerting system deployed into your environment, with a runbook and a team trained to operate it.
Retained advisory
Ongoing
A standing technical reviewer for in-house modeling work, and a named contact for the questions that arrive later.
Tools are chosen per problem. Where an existing package fits, we use it. Where it does not, we build the missing piece and hand it over with the rest.
Bring the problem.
The specification is part of the work. A first conversation costs nothing and usually saves a month.