Platforms · Google Cloud Engineering
We design, build and run Google Cloud environments for Canadian organisations: Cloud Run and GKE services, BigQuery analytics, Vertex AI workloads, and identity, cost and observability controls that hold up in production.
You own the code and IP Weekly demos Reply within one business day
$oldevs deploy --env production
✓Build passed · 212 tests · 0 warnings
✓Migrations applied · postgres 16
✓API v2 healthy · 142ms p95
→Rolling out to 3 regions…
OlDevs is a full-stack technology studio in Vancouver, British Columbia that engineers on Google Cloud.We build container and serverless workloads on Cloud Run, Google Kubernetes Engine and Cloud Functions, store data in Cloud SQL, Firestore and Cloud Storage, and run analytics in BigQuery with machine learning on Vertex AI. We deploy into the Canadian regions in Montreal and Toronto when data residency matters, and we set up identity, budgets and observability from the first sprint. Clients own all code, accounts and IP, and the next step is to request a quote.
Key facts
Platform expertise
We containerise services and run them on Cloud Run for request driven workloads, or on GKE when you need cluster level control. That includes build pipelines, autoscaling settings, rollout strategies, health checks and sensible resource limits.
Small pieces of logic belong in Cloud Functions, wired to Pub/Sub topics, Cloud Storage events, Eventarc triggers and Cloud Scheduler. We keep each function single purpose, add retries and dead letter handling, and document every trigger and its payload.
We pick storage by workload: Cloud SQL for relational data, Firestore for document data and fast client sync, Cloud Storage for files and backups. Backups, restore drills, lifecycle rules and access policies are part of the build, not an afterthought.
BigQuery is the reason many teams choose Google Cloud. We model datasets, load them with Dataflow, Datastream or scheduled queries, control cost with partitioning and clustering, and publish reporting through Looker Studio or your own front end.
Vertex AI covers training, tuning, a model registry, batch and online prediction, and access to Google's Gemini models. We connect it to your BigQuery data, add evaluation and monitoring, and keep prompts, features and model versions under version control.
We set up IAM roles by least privilege, Workload Identity Federation instead of long lived keys, Secret Manager for credentials, budgets and alerts for spend, and Cloud Logging, Monitoring and Trace so failures are visible before users report them.
Process
Assess
We inventory workloads, data volumes, dependencies and residency rules, then agree what belongs on Google Cloud and what does not.
Architect
You get a written design: projects, folders, networking, IAM model, environments, storage choices, budgets and the services each workload will use.
Build
We deliver in weekly increments with a working demo every week, infrastructure as code, and pull requests you can review as we go.
Migrate
Data and traffic move in rehearsed steps with rollback ready. We test restores, run in parallel where sensible, and cut over on a planned window.
Operate and hand over
Dashboards, alerts, runbooks and cost reviews go live with the system, and your team is trained to run it without us if that is the goal.
How we work
BigQuery gives you a serverless warehouse with no cluster to size, so analytics teams can query large datasets without an infrastructure project first.
Cloud Run runs a container, scales it with traffic and scales it back to zero. It removes most of the day to day work of keeping a platform alive.
Vertex AI sits beside your data instead of on the other side of an export. Training, evaluation, prediction and Gemini model access live in one place.
Regions in Montreal and Toronto let you keep regulated workloads and backups in Canada, with organisation policy limiting where resources can be created.
Serverless products bill by use and scale down when nobody is asking. Campaigns, registrations and reporting peaks stop being a capacity conversation.
If your organisation runs on Google Workspace, Analytics and Ads, identity and data exports line up with far less glue code to write and maintain.
Who it's for
Engineering leaders who need Cloud Run or GKE services, a BigQuery warehouse and an IAM model their security team will sign off, delivered without disrupting the platform already in production.
Programs with residency, accessibility and procurement obligations. We deploy into Canadian regions, meet WCAG 2.2 AA on the interfaces we build, and can work bilingually in English and French.
Multi-location operators who need one data model across every branch: location level reporting in BigQuery, shared services on Cloud Run, and per-location access controlled through IAM.
Founders who want to ship without buying a platform team. Cloud Run and Firestore keep the early stack small, and the architecture leaves room for BigQuery and Vertex AI when the data arrives.
Since 2014
Years building and running cloud systems
Weekly
Cadence of working demos during a build
2
Canadian regions used for residency work
FAQ
Day to day we build on Cloud Run, Google Kubernetes Engine and Cloud Functions, store data in Cloud SQL, Firestore and Cloud Storage, and run analytics in BigQuery with machine learning on Vertex AI. Around that we configure IAM, Secret Manager, budgets and Cloud Monitoring so the environment is safe to operate.
Yes. Google Cloud has Canadian regions in Montreal and Toronto, and we deploy resources into them when residency matters. We also set organisation policy to restrict resource locations, keep backups in Canadian regions, and record in writing which services are regional and which are global, current at the time of writing.
Cloud Run suits request driven and event driven services: you ship a container, it scales to zero, and there is no cluster to run. GKE earns its keep when you need custom networking, GPUs, operators, sidecars or long running workloads with specific scheduling. Many teams run both, and we help you draw the line.
Cost work starts in the architecture. We size services and choose serverless where traffic is spiky, set partitioning and clustering in BigQuery so queries read less data, apply lifecycle rules to Cloud Storage, and add budgets, alerts and resource labels so spend can be traced back to a team or workload.
Yes. We start with an inventory of services, data volumes and dependencies, then propose a route for each workload: rehost, containerise for Cloud Run or GKE, or rebuild. Data moves with tools such as Database Migration Service, Datastream or Storage Transfer Service, and we rehearse the cutover before the real one.
You do. Projects live in your own Google Cloud organisation and billing account, and code lives in your repositories. Clients own all code, designs, accounts and IP. We work through your identity provider with named accounts, and at the end of an engagement we hand over documentation and remove our access.
Send us the outline of your workload, your data residency needs and any deadlines. We reply to every enquiry within one business day, then book a call in your time zone. After a short discovery we send a written scope with milestones and a clear set of deliverables. Request a quote to begin.
Let’s connect
Tell us what you’re building. We’ll reply within one business day with next steps and a tailored quote — no obligation.