You deal with one party, agree one written scope and judge the result against acceptance criteria you helped write. Here is what happens at each step, who does the work and how your systems and data are treated.
Not every engagement uses every step. A well-defined sprint can go from qualification call to delivery in days.
You describe the outcome, the system and the constraints — through the contact form or by email. We reply with questions or a suggested next step, not a sales sequence.
A short call with the person who owns the outcome. We establish the business consequence, the deadline, who decides, what access is possible and whether we are the right people for it. Sometimes we are not, and we say so.
The scope names the deliverable, the acceptance criteria, what is excluded, who does the work and the fee. Nothing starts until both sides have agreed it in writing.
Uncertain problems start with a fixed-fee readiness review. Well-defined ones — a clear contract, an agreed benchmark, written acceptance tests — are quoted directly.
Work happens in your repositories and environments wherever possible, in small increments demonstrated against the acceptance scenarios rather than reported in status decks.
Agreed scenarios are run in the agreed environment. The result shows what works, what fails safely and what remains out of scope. Your owner signs it off.
Source, tests, runbooks and decision records go to the team that will own the system, with a walkthrough and time for questions.
Ongoing support or a next workflow is a separate decision, scoped on its own merits — never an assumed subscription attached to the first project.
Every engagement is led by the founder, who owns the architecture, the scope and the acceptance of the result. Smaller engagements are usually delivered entirely by the founder.
When a project needs capacity or a specialism we do not have in-house — a particular platform, a second engineer for a deadline — we bring in vetted specialists. They are named in the statement of work, work to the same acceptance criteria and confidentiality terms, and remain Neul Labs’ responsibility. You are never asked to assemble or coordinate suppliers yourself.
We also assess existing software before writing new code: commercial platforms, open-source components and our own public projects are all options, chosen on their merits for your situation.
Every measured claim states its method and conditions. A baseline is taken before anything changes, the same measurement is repeated afterwards, and the limitations are written down. The engineering evidence method describes the approach, and our case studies follow it.
Sometimes the work uncovers that an assumption was false. When that happens we stop, explain what was found and agree any change in writing. Scope is never expanded silently, and a finding that changes the plan is documented as part of what you receive.
Neul Labs is accountable for the whole scope. The founder leads architecture and acceptance on every engagement. Where a project needs specialist capacity, the specialist is named in the statement of work and works under the same scope, standards and confidentiality terms.
Yes, when that is the honest answer. An existing platform configured well, a small integration, ordinary deterministic software or no project at all are all legitimate recommendations, and we would rather make them early.
The least that does the job. We start with synthetic data and non-production environments, ask for read access before write access and agree how any sensitive data will be handled before touching it.
No. Our public projects are options to assess on their merits under their own licences. Many engagements use none of them, and adopting them is never a condition of working with us.
By demonstrating acceptance scenarios as they pass, in your environment, plus a short written note of decisions and open questions. If something is at risk, you hear about it when we know, not at the end.
Tell us the outcome, the system and the deadline. The first reply will be a question or a next step, not a proposal.
Scope a project