ResearchDefence and government work is shared on request.
Small teams working quietly on serious problems, with one difference from the usual advanced projects outfit – the door is open. We have gone looking for capability since day one, because the scale of what we are building demands it. If you are a builder who feels stuck, or a small company that cannot see the path, walk in. You will get straight talk and people who will work the problem with you.
Research is paid for by our delivery work, so nothing ships before it is ready and a weak idea costs nothing to kill. Venture capital has a place too – when a build spins out and needs to grow fast or reach new markets, we use it deliberately. The mission sets the pace, not the funding.
Part of our method exists to attack our own work. We try to break what we built, then fix what broke, before we deliver. Some teams put the plane in the sky and work on the engines later. We do not, and our tooling keeps us closer to their speed than you would expect.
Need
The same need keeps surfacing, and the workaround is costing someone.
One observation is an anecdote. The gate opens on a pattern with a cost we can measure.
Definition
The problem is written down precisely, before anyone is allowed to fall for an answer.
It covers who it hurts, what success looks like, and what conditions it must survive. How it will be built is deliberately left out.
Review
A formal review with accountable people in the room. Plenty stops here.
That is the system working. Names attach to the decision, including the decision to stop.
Build
Development under security discipline from the first commit.
We start from a hardened baseline, settle who can see what before the first feature, and bring in partner capability where it sharpens the result.
Validation
Judged against the original need, by people who did not build it.
It advances when it passes. The date follows the evidence, every time.
Four disciplines every Research build runs under, whether it is a board or a codebase. Open any of them for the technical position.
Hardware & software, one discipline
The bench gets the same rigour as the repository.
Research builds physical things as well as code. Component sourcing is treated as supply chain, firmware ships under the same pipeline controls as software, and test rigs produce evidence you can put in front of someone.
Data architecture & governance for regulated environments
In a regulated environment, the data model is the compliance position.
We classify the data before we model it, and decide which organisations and people can see what before the schema is drawn. Every boundary the data crosses is treated as a control that has to be justified.
Assurance & evaluation
Deploying is a conclusion the evidence reaches.
Automated evaluation runs through the whole build. It tests performance on the cases that matter, behaviour on the edge cases that hurt, and regression against everything that has failed before.
Security architecture & adversarial design
The threat model is written before the architecture is.
We decide what an attacker would go for and what it would cost them before the first design decision is made. The hardened baseline, the identity model and the logging are part of the build, not a review it has to survive at the end.
Research
Salus is the drafting layer between what a provider already knows and the care plan it owes. The first product out of 308 Research, built with Vox Novus under PartnershipDNA and now in use with Australian home care providers — what the five gates produce when a need keeps surfacing and nobody sells the answer.
Australian home care providers already hold everything a care plan needs. Salus pulls it into one client picture and an auditable draft in under 30 minutes. By hand that takes three to five hours, and detail still goes missing. A care manager reviews, edits and approves.
Visit Salus website
