A two-person studio for embedded product iteration, with GCP DevOps wired in from day one. Vertex AI, Gemini Enterprise, multi-model orchestration: built into your codebase, not handed over a fence.
"We don't ship code over the fence."
We work inside your repo, on your branch protections, in your standup. Pull requests on day three; design docs that match what's actually deployed. The engagement ends when your team owns the result.
"From day one. Not retrofitted."
Terraform from first commit. Cloud Build pipelines, Cloud Run services, GKE Autopilot clusters, Pub/Sub fan-out, BigQuery sinks. We don't bolt CI onto a six-month-old repo. We set it up before the first feature lands.
"Multi-model. Vertex AI. Gemini Enterprise agents."
Deterministic routing across Gemini 3, Claude Opus 4, and GPT-5, with budget caps in milliseconds and dollars. Agent Engine on Vertex AI or GPU-backed inference on GKE Autopilot, deployed on your VPC, wired to existing IAM. Eval harness running on day one.
The biggest single lever on Vertex (50% off, same model) is running the workload as a batch job. Most AI cost discourse stops at per-token price. The more useful question is whether each workload actually needs sub-minute latency.
We ran this on a pipeline processing 20,000+ PDFs a month with Gemini on Vertex: two paths, one prompt, one model. Real-time for ad-hoc requests; batch for everything else. A Cloud Workflow builds the JSONL, submits to Vertex Batch Prediction, polls until done, writes to both stores, and falls back to the real-time API for any failed rows. Orchestration on Cloud Workflows plus Cloud Run Jobs costs a few dollars a month.
Full breakdown on LinkedIn ↗
Eleven years embedded inside one of the region's largest privately held payment processors, first as CTO and then as CIO, building production systems across payments, cloud infrastructure, enterprise data, pricing intelligence, and applied AI. Vertex AI, AlloyDB, BigQuery, GCP, and multi-model orchestration were not retrofitted as talking points. They were built into the work.
SaaS 971 is built around a simple belief: small, senior teams can still ship serious software.
Derek builds production systems for companies where the gap between business need and engineering capacity is expensive. His work spans payments infrastructure, pricing intelligence, data platforms, AI-assisted workflows, and cloud-native applications on GCP.
The common thread is practical: turn unclear, manual, high-friction processes into software that operators, executives, and customers can actually use. No oversized roadmap. No unnecessary vendor layer. Just senior technical judgment, business context, and production execution.