Patterns
The solutions we've shipped — repeatable, anonymized.
We can't always name our customers. We can describe exactly how we solved their problem, what it took, and what it cost. If your problem rhymes with one of these, we can probably help.
-
Cross-sector
May 2026
How we got autoscaling, canary, and AI-powered ops on a 60-container stack without taking on a platform team
How we moved a Canadian operator's 60-container production stack off plain compose without making the conventional jump to a heavyweight orchestrator. Workloads stayed on the operator's own hosts; the orchestration control plane is operated as a managed service. No platform-engineering hire, no multi-component cluster, no foreign-cloud residency posture.
Typical duration
Under 12 hours for a 50–80 container stack when the manifest configuration is already known
-
Cross-sector
May 2026
How we consolidated four products' billing onto a self-hosted core without paying a percent of every transaction
How we replaced a per-product, percent-of-revenue billing-as-a-service stack with a self-hosted billing core for a multi-product Canadian platform operator. Each product kept its own pricing logic and its own UI; the customer relationship, the audit ledger, and the residency posture moved into the operator's perimeter.
Typical duration
12–16 weeks for a four-product portfolio
-
Government
May 2026
Building integration that survives a four-hour satellite outage
How we designed offline-first integration for a Northern anchor institution: local-first state, queue-and-replay sync, idempotent integrations. The field worker never knows the link dropped, and the back-office systems reconcile cleanly when it returns.
Typical duration
10–14 weeks for the first deployment
-
Energy
May 2026
Building a regulator-grade voltage audit trail a regulator will accept months later
How we shipped independent voltage-presence witnessing at the customer end of a distribution network — so when the regulator audits a service-band claim six months after the fact, the evidence stands without depending on the distribution operator's own systems.
Typical duration
12–18 weeks per deployment phase
-
Government
Apr 2026
How we shipped AI workflow automation inside a Canadian municipality's tenant
Natural-language workflow building (Kacha) deployed inside the customer's tenant — no data leaves the perimeter, no license per workflow, no foreign cloud. From scoping to first workflow in production in under ten weeks.
Typical duration
8–12 weeks from scope to production
-
Telecom
Apr 2026
Agentic AI on a thirty-year-old OSS, without ripping it out
How we wired an agent runtime into a telecom operator's legacy customer system using a custom protocol connector and reusable skill files — turning a read-only audit nightmare into a queryable, governable AI workspace for operations teams.
Typical duration
8–12 weeks for a production read-access pilot
-
Cross-sector
Apr 2026
Replacing legacy field-service applications without disrupting operations
How we migrate operationally critical field-service applications onto the operator's existing enterprise platform — preserving the workflow, replacing the lock-in. Worked example: a continental field-service replacement that ran every shift through the cutover without missing a dispatch.
Typical duration
10–14 weeks end-to-end