You have carried production for a multi-tenant B2B SaaS with paying customers

You have designed and enforced least-privilege access for a real engineering team.

An identity provider, role design, and the removal of access from people who had it.

Deep GCP — Cloud Run, Cloud SQL, VPC, IAM, Secret Manager, workload identity

Terraform in production, including codifying an existing estate.

Extensive experience operating on GITLAB.

CI/CD you built and made the only path, including runner and secrets hardening.

PostgreSQL operations — pooling, replicas, failover, a restore you ran.

You write code — enough Python or C# to open the merge request yourself.

Ability to work in cross-functional teams.

High proficiency in English communication.

The path to production: The pipeline is the only way anything reaches a client: merge request, review, green tests, staged rollout, rollback that has been used. Migrations managed across environments. Hardened runners. Delivery that does not depend on who is awake.

Identity & access: One identity provider across GitLab, GCP, production data, and internal tools. Role-based access designed around what each engineer does, reviewed on a schedule, time-boxed elevation for the exception. Joining and leaving are checklists that run in under an hour.

Secrets, keys & devices: Secret Manager, masked CI variables, workload identity instead of long-lived keys, rotation with an owner. A minimum standard for the personal machines the team works from, and a way to verify it. Secrets never live on a laptop.

Tenant operations: Onboarding a client is a pipeline which includes provisioning, configuration, seeding, and a segregation model for deployment, data, and audit that is built in rather than bolted on.

Infrastructure as code: Terraform across the whole estate; environments that can be destroyed and recreated. Private data tier, controlled egress, edge protection, organisation policy that makes the safe thing the default.

Audit & detection: Who touched which system and who opened which patient record — tamper-evident, queryable, written to a sink you do not control. Alerts on what should never happen.

Observability & on-call: Tracing across .NET and Python so a booking is followable end to end. SLOs per clinical journey, alerts that page on a broken journey, a blameless incident process with post-mortems people read.

PostgreSQL & real-time: Pooling that survives Cloud Run scale-out, replicas, failover, restore drills actually run. WebSocket sessions that last a whole appointment, kept alive across scaling.

Healthcare or other regulated SaaS.

Understanding and experience of HIPAA, NPHIES, PDPL, and ISO 27001.

Kafka or Pub/Sub — we move to an event bus after the current direct-HTTP phase.

You have carried production for a multi-tenant B2B SaaS with paying customers

You have designed and enforced least-privilege access for a real engineering team.

An identity provider, role design, and the removal of access from people who had it.

Deep GCP — Cloud Run, Cloud SQL, VPC, IAM, Secret Manager, workload identity

Terraform in production, including codifying an existing estate.

Extensive experience operating on GITLAB.

CI/CD you built and made the only path, including runner and secrets hardening.

PostgreSQL operations — pooling, replicas, failover, a restore you ran.

You write code — enough Python or C# to open the merge request yourself.

Ability to work in cross-functional teams.

High proficiency in English communication.

The path to production: The pipeline is the only way anything reaches a client: merge request, review, green tests, staged rollout, rollback that has been used. Migrations managed across environments. Hardened runners. Delivery that does not depend on who is awake.

Identity & access: One identity provider across GitLab, GCP, production data, and internal tools. Role-based access designed around what each engineer does, reviewed on a schedule, time-boxed elevation for the exception. Joining and leaving are checklists that run in under an hour.

Secrets, keys & devices: Secret Manager, masked CI variables, workload identity instead of long-lived keys, rotation with an owner. A minimum standard for the personal machines the team works from, and a way to verify it. Secrets never live on a laptop.

Tenant operations: Onboarding a client is a pipeline which includes provisioning, configuration, seeding, and a segregation model for deployment, data, and audit that is built in rather than bolted on.

Infrastructure as code: Terraform across the whole estate; environments that can be destroyed and recreated. Private data tier, controlled egress, edge protection, organisation policy that makes the safe thing the default.

Audit & detection: Who touched which system and who opened which patient record — tamper-evident, queryable, written to a sink you do not control. Alerts on what should never happen.

Observability & on-call: Tracing across .NET and Python so a booking is followable end to end. SLOs per clinical journey, alerts that page on a broken journey, a blameless incident process with post-mortems people read.

PostgreSQL & real-time: Pooling that survives Cloud Run scale-out, replicas, failover, restore drills actually run. WebSocket sessions that last a whole appointment, kept alive across scaling.

Healthcare or other regulated SaaS.

Understanding and experience of HIPAA, NPHIES, PDPL, and ISO 27001.

Kafka or Pub/Sub — we move to an event bus after the current direct-HTTP phase.