Consulting & DevOps

What this is

The engineering work that isn't a project — it's a decision, an audit, or a presence on the team.

A startup hits Series A and the angel CTO is out of their depth. A scale-up's deploys break every Friday and nobody knows why. A VC wants a tech due-diligence read before writing a cheque. A mid-size SaaS has Postgres on fire and no idea which query is responsible.

We solve these the same way we ship products: time-boxed, with a concrete deliverable.

Architecture review

Two weeks. One senior engineer. One concrete report.

We come in, we read the code, we run it, we measure it. We talk to the engineers who built it about the parts that scare them. We come back with:

1. A one-page executive summary your board can read.
2. A ranked list of risks — critical, important, can-wait.
3. For each risk, an estimated fix cost and the expected lifespan if you don't fix it.
4. An honest read on the team's technical debt to delivery-velocity ratio.

We don't sugar-coat. If we think the system needs a rebuild, we'll say so — and explain why we think so.

Used for: VC due diligence, post-acquisition reviews, internal "should we rebuild this?" discussions.

Infrastructure and CI/CD setup

Two-to-four weeks. Outcome: predictable deploys, clean environments, observability you can act on.

You have a system. It works. Deploys are stressful, environments are inconsistent, secrets live in a Notion doc. We bring it to a state where:

• A new engineer can run the whole stack locally in under an hour.
• git push to main triggers tests, builds, and deploys staging.
• Production deploys are a one-click promotion with rollback in under two minutes.
• Logs, metrics and traces are correlated and queryable.
• Secrets live in a real secret store, not anyone's .env.

We use whatever your team will inherit comfortably — GitHub Actions, Docker Compose, OpenTelemetry, Grafana. We pick stable boring tools that your engineers will still be productive with in three years.

Performance audits

One-to-three weeks depending on system size.

Your system is slow. You think you know which part. You're probably half-right.

We profile end-to-end: frontend Core Web Vitals, API latency distribution, database query plans, background job throughput. We find the actual bottlenecks — not the ones people complain about. We fix the top five and hand over dashboards that will tell you when number six becomes the new top one.

Typical wins on inherited systems: 60-80% reduction in p95 API latency, 2-5x improvement in throughput on heavy endpoints, Core Web Vitals from "needs improvement" to "good" on all key pages.

Fractional CTO retainer

Eight to sixteen hours a month, for as long as you need it.

You're a non-technical founder past your first ten customers, or a technical founder who needs an external sanity check. You're not ready for a full-time CTO hire — but the early-stage CTO-decisions you make in the next six months will haunt the company for five years.

In a retainer we:

• Review code weekly. Catch architectural drift before it becomes debt.
• Sit in on hiring loops. Help write the engineering JD, screen the senior candidates.
• Make vendor decisions with you. Database, observability, CI, hosting — the calls that lock you in for years.
• Pair on the hardest technical problems. Not just whiteboard advice — we'll write code if the team is stuck.

We retain a small number of these engagements at a time. They are not a side hustle; they get real attention.

What we don't do

• One-call advisory. The shortest engagement we offer is two weeks.
• "Just give us your honest opinion in a Slack message." Honest opinions come with documents that justify them.
• Pure conference speaking or content marketing under a consulting label.
• Replace your existing CTO on day one. We work alongside them — we don't unseat people.