47HQ
All Services

08 · AI Infrastructure

DevOps & Infrastructure

Production-grade infra your team can extend, scale, and on-call against.

ship-pipeline · prod
LIVE
Release · main @ a7f3c92
CI gated by evals + SLOs
green
build1m 12s
unit48s
evals2m 04s47/47
canary 5%6mSLO ✓
promote 100%now
Lead time
11 min
commit→prod
Change-fail
1.8%
↓ 6pt
MTTR
4 min
auto-rollback
CI gated by evals and SLOs.
47hq
Duration
3–6 weeks
Team
1 principal + 1 engineer
Starts in
Kick-off within 1–2 weeks of SOW
Investment
Fixed fee · $50k–$110k

Overview

What you get

IaC, CI/CD for prompts and retrieval, autoscaling tuned to your latency and cost targets, and a security posture review.

The problem

Why teams call us

  • AI services live in a hand-rolled deploy that only one person understands.
  • Prompts and retrieval indices are versioned in Notion, not in CI.
  • Autoscaling either over-provisions cost or under-provisions latency.

Approach

How we work

  • Treat prompts, retrieval, and models as deploy artifacts in CI.
  • Codify infra so any engineer can extend it, not just the original author.
  • Tune autoscaling against real production traffic, not a synthetic load test.

Process

Week by week.

  1. 01 · Week 1

    Audit

    Current infra, security posture, secrets, network.

  2. 02 · Week 2–3

    Codify

    IaC, CI/CD with prompt + retrieval versioning.

  3. 03 · Week 4–5

    Tune

    Autoscaling against real traffic, cost + latency targets.

  4. 04 · Week 6

    Handoff

    Runbooks, on-call rotation guide, security review.

You're a fit if
  • AI workloads outgrowing a hand-rolled deployment
  • Need CI/CD that respects prompts, retrieval, and models as artifacts
  • On-call rotation that wants better signal
Probably not a fit if
  • Pre-production prototypes still finding product-market fit
  • Teams without anyone on call
  • Orgs unwilling to adopt IaC

Deliverables

Everything we ship

  • 01IaC for inference + retrieval services in your cloud
  • 02CI/CD with prompt + retrieval versioning
  • 03Autoscaling tuned to your latency and cost targets
  • 04Secrets, network, and access posture review
  • 05On-call runbook tailored to your stack

Outcomes

What you walk away with.

Minutes
deploy time, not days
−30%
infra cost via right-sized autoscaling
1 runbook
any on-call engineer can act on

Production infra your team can extend, scale, and on-call against — with deploy times measured in minutes, not days.

Tooling

Stack we ship against

Model- and infra-agnostic. We adapt to your stack, not the other way around.

TerraformPulumiGitHub ActionsAWSGCPCloudflareFly.ioDatadog

FAQ

Real questions, technically answered.

Terraform or Pulumi?
Whichever your team already uses. We meet your stack, not the other way around.
Do you set up Datadog / Grafana / etc?
Yes. We extend whatever observability stack you already pay for.
Can this include SOC 2 prep?
We don't run the audit, but we leave you ready for one — secrets, access logging, network segmentation.

Next step

Ready to scope DevOps & Infrastructure?

Book a discovery call. We'll confirm fit, sequence the engagement, and have a Statement of Work in your inbox within a week.

Refundable if we're not a fitWritten diagnostic in 48 hoursSession run by a founder, not a sales rep