Skip to content

Custom Engineering

Service ownership

Owner: professional-services (ps-pm@clouddigit.ai) — Status: GA — Last audited: 2026-05-11

Bespoke engineering work — Terraform modules, integrations, automation, glue.

What it is

When the off-the-shelf services don't quite line up with what you need to build, Custom Engineering bridges the gap. Typical work:

  • Custom Terraform modules for your in-house IaC pattern
  • CI/CD integrations with your existing pipeline (Jenkins, GitLab, Bitbucket Pipelines)
  • Service-mesh integrations for K8s
  • Identity federation (SAML / OIDC) with your IdP — Keycloak, Azure AD, Okta
  • Custom backup / DR orchestration beyond what BaaS / DRaaS provide
  • Migration tooling for unusual source systems
  • Cost / chargeback automation that pulls from billing API and pushes to your finance system

Engagement shape

T&M, with a defined Statement of Work that lists deliverables and acceptance criteria. Most engagements are 2–8 weeks.

What you get back

  • Source code (yours; we don't keep IP)
  • Tests
  • Deployment scripts
  • Documentation
  • A walkthrough session with your team

Pricing

T&M, by senior engineer-day. See Pricing.

Operate this service

CD engineering team for one-off custom development — integrations, automation, custom features.

What fits

  • Custom integration between CD service and a customer-specific system
  • Automation scripts / IaC modules tailored to customer
  • Performance work (kernel tuning, custom NIC drivers, low-latency)
  • Custom dashboards / reports
  • One-off compliance tooling

What doesn't fit

  • Long-term application development (use a system integrator)
  • Customer's product work (out of scope)
  • Open-ended R&D (need scoped deliverables)

Engagement model

  • T&M for exploration phases
  • Fixed-price for well-scoped deliverables
  • 4-12 week typical engagements

IAM

Role Can do
custom-eng.viewer View engagement deliverables
custom-eng.requester Submit requests
custom-eng.admin Sign off engagement scope and acceptance

CD engineers get engagement-scoped access; output handed over to customer at completion.

Deliverables

Standard: - Working code (in customer's repo) - Documentation - Tests - Deployment instructions - Runbook for ongoing operation

Ownership transfers at acceptance; CD doesn't maintain post-engagement unless retained separately.

Pricing

  • Senior engineer day rate (negotiable for multi-month)
  • Discounts for fixed-price deliverables
  • Materials at cost

Request workflow

bash cd professional custom-eng request \ --title "Custom S3 → Splunk pipeline" \ --description "..." \ --budget-cap 40h

CD review: - Scoping conversation (1-2 sessions) - Feasibility assessment - Proposal (timeline, cost) - Customer approves; engagement starts

Sprint cadence

For multi-week engagements: 1-2 week sprints with demos. Short engagements: daily standups.

Acceptance criteria

Pre-defined at scoping: - Functional requirements (clearly defined "done") - Non-functional (performance, reliability) - Documentation deliverables - Test coverage

Acceptance is a checklist; ambiguity is the enemy.

Code ownership

Customer's repo, customer's branch policies. CD engineer commits with their CD principal; PRs reviewed by customer team.

Open source contributions: tracked separately if delivered output goes upstream.

Knowledge transfer

End-of-engagement: - Code walkthrough - Architecture document - Operations runbook - Q&A session - Handover sign-off

Customer team should be able to maintain post-handover.

Hypercare

2-week hypercare period post-engagement: - Bug fixes for defects in delivered work - No new features - Light support

Beyond: separate engagement.

Scope creep

Customer adds requirements mid-engagement: - Document as change request - Estimate impact (cost, timeline) - Customer approves change or defers

Engagements without change control overrun.

Acceptance criteria disputed

CD delivers; customer says "not what we asked for": - Compare against signed acceptance criteria - If CD missed: fix as part of engagement - If criteria was ambiguous: extend scope (negotiated) - If customer wants different thing: change request

Knowledge transfer didn't take

Post-engagement, customer team can't operate: - Hypercare extended (paid) - Additional training engagement - Re-engagement to refactor for clarity

Prevention: continuous KT through engagement, not just at end.

Delivered code has bugs

Within hypercare: fix at no additional cost.

After hypercare: bug-fix engagement; customer pays for fix time.

Estimate exceeded

Engagement taking longer than planned: - Re-estimate with cause - Customer decides: continue (more cost), reduce scope, or stop - Document cause for future engagements

CD engineer assignment changes

Mid-engagement, a CD engineer rolls off: - Replacement engineer onboards (1-3 days ramp) - Engagement continues; some loss of momentum - For long engagements: pre-designated backup engineer

Customer team underpowered

CD doing 90% of work, customer team minimal: - Risk: customer team can't operate handover - Surface concern; re-balance involvement - May need to extend engagement for proper KT