Product Engineering
An engineering partner for the whole product lifecycle — discovery, architecture, build, and the unglamorous work of keeping it healthy after launch.
Explore this serviceA SaaS product is easy to start and expensive to correct. Tenancy, permissions, billing and data isolation are decided in the first month and questioned in the first serious security review. We build those foundations properly, then ship the product on top of them.
Most SaaS rewrites are not caused by scale. They are caused by a tenancy model that cannot express what a customer now needs — a user in two organisations, an enterprise wanting SSO and a separate data region, a role that does not fit the three that were hardcoded.
We decide the isolation model deliberately: shared schema with row-level security, schema-per-tenant, or database-per-tenant, chosen against your actual compliance and scale requirements. Every query is scoped by construction, so a missing filter is a type error rather than a data leak.
Roles are resource-and-action based from the start, because 'admin, member, viewer' will not survive your first enterprise conversation. Billing is wired to entitlements and reconciled against the provider, so what a customer pays for and what they can do never drift apart.
The situations teams are usually in when they bring us this problem.
A tenant id added to tables after launch leaves gaps. One missing filter is a cross-tenant leak and an incident report.
Entitlements drift from subscriptions, so customers pay for what they cannot use — or use what they never paid for.
Every enterprise deal asks for a fourth. Resource-action permissions cost little up front and save the rewrite.
Audit trails, isolation and access control get asked about before the contract. Retrofitting them stalls the deal.
The same sequence on every engagement, so there are no surprises in week three.
Tenancy, isolation, identity, permissions and billing model — written down with their trade-offs before any product feature is built.
Auth, organisations, roles, subscriptions and audit logging as one coherent layer the product sits on.
Features on top of a foundation that already handles who can see what and who is paying for it.
The documentation, logging and controls your first enterprise buyer's security review will ask for.
Concrete deliverables — all of it yours to keep, run, and hand to another team.
The isolation model chosen against your compliance needs, with scoping enforced by construction.
Authentication, organisations, invitations, and resource-action permissions ready for SSO.
Subscriptions, plans, entitlements, proration and dunning, reconciled against the provider.
Impersonation with an audit trail, plan changes, and the operational tools support needs.
Who did what, when — plus the metrics and traces to explain a slow tenant.
Chosen for the problem, not for novelty.
We are not tied to any of these. If your team already runs something that works, we build with it.
Auth, tenancy and permissions as one layer, so no feature has to reinvent access control.
Row-level security and query-time scoping, so isolation does not depend on every developer remembering.
Entitlements derived from the subscription, reconciled continuously against the billing provider.
The questions that come up in almost every first conversation.
Typically three to five months for a product with real tenancy, billing and permissions. Roughly a third of that is the platform layer — and it is the third that determines whether year two is comfortable.
Yes, and we have done it several times. It is more work than doing it first, and it needs a careful migration with verification, but it is a well-understood exercise rather than a rewrite.
No — but the identity model should be able to accept it without a rewrite. We design for that, then implement SSO when a deal actually requires it.
Bring the problem, not a spec. We will tell you honestly whether we are the right team and outline a concrete first step — no obligation, no sales theatre.