DevOps Consulting

The shortcut: Stop selling "CI/CD pipelines." Sell unblocked velocity — measured in deploys per week and minutes-to-recover from incidents. The people writing the check don't know what Terraform is, but they know their developers were stuck for three days last sprint.

Industry: Software & Tech | Investment level: Small — $2,000-$10,000 | Time to launch: 4-8 weeks (one cert refresh + a paid discovery template + first audit gate the launch) Best for: A senior engineer or SRE comfortable in AWS or GCP who can read a client's GitHub Actions logs, sketch an architecture diagram on a Zoom call, and explain a blast radius to a non-technical CEO without making them panic. What you'll likely make: $3,000-$6,000 month 3, $8,000-$15,000 month 6, $15,000-$25,000 month 12. Math is in Section 4.

Market Opportunity

A senior DevOps engineer in the US costs $180K-$280K all-in once you load benefits, equity, and recruiter fees (Levels.fyi DevOps comp range). A 12-person Series A startup needs that work two days a week, not five — and can't make the math work on a full-time hire. That gap is the entire business.

The crowded part is "we do DevOps" agencies pitching $200/hour Kubernetes migrations to Fortune 500s. The quiet part is the 5-30 person SaaS startup whose three-engineer team ships once every two weeks because nobody has owned the build pipeline since the founding CTO got pulled into product. They'll pay $4K-$8K for a three-week audit-and-fix, then $2K-$4K/month to keep things tidy. None of them are Googling "DevOps consultant." They're complaining about it on Slack.

What anchors the conversation is DORA metrics — deployment frequency, lead time for changes, change failure rate, and mean time to recover. The 2023 DORA report benchmarks elite performers as deploying multiple times per day with a change failure rate under 5%; low performers ship once per month or less (2023 Accelerate State of DevOps Report). Walking a CEO through those four numbers is a concrete audit deliverable in plain English — and it justifies the next engagement.

Demand is not the problem. Scope discipline is.

Launch With AI

Pro section. AI doesn't replace the on-call work — it cuts the parts that drained you (writing the DORA audit deliverable, drafting Terraform module boilerplate, generating GitHub Actions YAML, building monthly retainer reports). Spend the saved time on what AI can't do: standing up the right Terraform module structure, validating the deploy pipeline against the client's actual production data, and walking the CTO through the cutover plan at midnight when something actually breaks.

The trap most first-year DevOps consultants fall into: they think AI will research AWS service limits or Kubernetes resource quotas for them. It will hallucinate quota values + KB numbers that don't exist. AI for the writing tail (audits, SOWs, retainer reports) AND for IaC scaffolding boilerplate is fine. But every Terraform plan + every GitHub Actions workflow + every kubectl apply gets your eyes on it before it touches production.

Important up-front: AI cannot run the cutover, refuse the fixed-price K8s migration prospect, or restore from a Terraform state file at 2am. It will also confidently generate "zero downtime" guarantees you can't deliver solo. You own every commitment, every architecture decision, every incident response; AI scales the writing around them.

AI Tools You'll Use

Tool	Price	What it does
ChatGPT Plus	$20/mo	DORA audit reports, SOW templates, sales follow-ups, retainer summaries
Claude Pro	$20/mo	Long-context CI/CD log analysis, IaC review, multi-cloud doc synthesis
Cursor	$20/mo	Terraform/Pulumi modules, GitHub Actions YAML, runbook code, kubectl manifests
Datadog / Grafana built-in AI	included with paid tier	Anomaly detection, alert correlation, incident pattern recognition
Loom AI	free	DORA walkthroughs, incident postmortems, retainer monthly reports

The Workflow

DORA audit deliverable (Claude long-context, ~3 hours/client). The 2-week audit is your wedge. Paste the client's GitHub Actions logs + the IaC repo + the cloud bill + the on-call setup into Claude:

"Below is the audit data for [client, N engineers, Series A SaaS]. Includes: GitHub Actions workflow YAML + last 90 days of run history, Terraform repo structure, AWS Cost Explorer last 6 months, PagerDuty incident log. Generate the 10-page DORA audit: (a) score the 4 DORA metrics — deployment frequency (current vs elite), lead time for changes (PR open to prod), change failure rate (% of deploys causing rollback or incident), MTTR (time to recover from production incident), (b) the 8 specific things broken (named: 'GitHub Actions runs sequential when they could parallel — 12 min wasted per deploy,' 'no IaC for the staging RDS — drift risk,' 'on-call rotation has 1 person, no runbook'), (c) the 12-week stabilization plan (week-by-week, what's fixed first, expected DORA lift), (d) the CFO summary (1 paragraph, the headline numbers, the cost vs ROI of the recommended fixes). NEVER hallucinate AWS quota values or GitHub Actions runner pricing. Tone: senior SRE advisor. Output as Notion-ready 10-pager."

Verify every quota + KB cite against AWS docs before sending. Claude long-context turns a 12-hour audit into a 3-hour audit — that's billable hours back to the technical work.
Reusable Terraform module library (Cursor + Claude, ~6 hours one-time). Your module library is the asset that makes pipeline build-outs profitable. Open Cursor:

"I'm building a reusable Terraform module library for SMB SaaS clients on AWS. Generate: (a) VPC module (3 AZs, public + private + database subnets, NAT, IPv6 dual-stack, flow logs to S3 + CloudWatch), (b) ECS Fargate module (cluster + service + ALB target group + autoscaling + CloudWatch alarms), (c) RDS Postgres module (Multi-AZ + automated backups + parameter group + subnet group), (d) S3 + CloudFront static-asset module (origin access identity + CORS + cache-control headers), (e) IAM Identity Center module (SSO with Okta SAML + permission sets per environment). Use AWS provider 5.x. Modules use semantic version tags. Code only with var files for client-specific overrides."

Build once. Client 4 onward, deploying the baseline takes 4 hours instead of 30. That's the unit-economics math for the $7K pipeline build-out.
GitHub Actions workflow + IaC PR review (Cursor + ChatGPT, ~60 min/build). For each new client, generate the GitHub Actions workflow. Paste:

"I'm building the GitHub Actions CI/CD for [client]. Stack: Node 20 + Postgres + ECS Fargate. Generate: (a) the .github/workflows/ci.yml that runs on every PR (lint + typecheck + unit tests + integration tests against ephemeral Postgres + Snyk security scan + image build with cache), (b) the deploy-staging.yml that runs on merge to main (build + tag image + push to ECR + Terraform apply staging + smoke test), (c) the deploy-prod.yml that runs on tagged release with manual approval gate (build + tag + push + Terraform plan with required approval + apply + smoke test + Datadog deploy marker), (d) the on-failure rollback workflow (revert to last green ECR tag + Terraform apply rollback). Output as 4 YAML files. Use github-actions-cache for build artifacts."

Verify every step in Cursor before committing. The 60-min review per workflow is what protects you from a deploy that bricks production.
Incident postmortem + retainer monthly report (Loom AI + ChatGPT, ~30 min/incident OR /client/month). Each incident becomes a postmortem that justifies the retainer. Paste:

"I run the fractional DevOps retainer for [client]. This week: (a) production deploy at [time] failed, rollback completed in [time], root cause was [paste], (b) we added a check to prevent recurrence (paste). For the postmortem, generate: (a) timeline (T-15 min through T+45), (b) root cause + contributing factors, (c) the 3 prevention items I'm implementing this week, (d) the 1-paragraph CTO-facing summary. NEVER assign blame to a specific engineer. Tone: senior SRE, blameless. Output as Notion-ready 1-pager + a 4-min Loom script the team can watch."

Loom AI auto-titles + transcribes. The blameless postmortem is what builds the cultural trust that keeps the retainer renewing year after year.
DORA self-assessment lead magnet (ChatGPT, ~60 min one-time). The 10-question Notion page is what converts cold LinkedIn outreach. Paste:

"I'm a DevOps consultant. Build a 10-question DORA self-assessment Notion page I can use as a free lead magnet on LinkedIn cold outreach: (a) 10 questions covering deployment frequency, lead time, change failure rate, MTTR, on-call setup, IaC coverage, observability maturity, security posture, secrets management, disaster recovery, (b) auto-scoring logic (per question: 1-4 scale → composite DORA performer level: low / medium / high / elite), (c) the 'where you score' results page (the score + the 3 highest-leverage fixes for that score band), (d) the 1-line CTA at the bottom ('want a 60-min audit walkthrough? book here'). Tone: senior SRE. NEVER 'unlock your potential' filler. Output as Notion-ready Markdown + auto-scoring formula."

Drop into Notion. Gate behind email. 10-15% of downloaders book a paid discovery audit — that single page drives 2-3 paid audits/month at zero ad spend.

Time Saved Per Week

Roughly 8-12 hours/week once your audit template, IaC library, and retainer reports are built:

DORA audit deliverable: 12 hours/client → 3 hours (Claude long-context)
Reusable Terraform: 30 hours/client → 4 hours (module library + Cursor)
GitHub Actions workflows: 4 hours/build → 60 min (Cursor)
Incident postmortems: 90 min/incident → 30 min (Loom AI + ChatGPT)
Retainer monthly reports: 60 min/client → 30 min (ChatGPT)

Trade that time for: the AWS DevOps Pro certification refresh, the conference talk you've been delaying, and the third fractional retainer client.

Total AI Stack Cost

Budget tier ($20/mo): ChatGPT Plus only. Cursor free for small Terraform; Datadog free trial; Loom AI free. Right for first 60 days while building the module library.
Full tier ($60/mo): ChatGPT Plus + Claude Pro + Cursor. Worth it once you sign anchor client 2 — Claude long-context on CI/CD log analysis alone saves 4+ hours per audit.
Compare: A junior SRE for IaC + audits + retainer reports runs $7,000-$12,000/mo. The full AI stack is one-hundredth that cost — and you keep eyes on every Terraform plan + every kubectl apply before production.

Cancel anything you don't open in a 7-day window. NEVER use AI for compliance research without verification against the actual cloud provider docs — hallucinated quotas are the kind of mistake that ends an engagement.

Your First Win

30 minutes from now your DORA self-assessment lead magnet is built. Open ChatGPT (free tier works). Paste:

"I'm a DevOps consultant targeting Series A SaaS startups (5-30 engineers). Build me a 10-question DORA self-assessment as a Notion-ready lead magnet for LinkedIn cold DMs. Cover: (a) the 10 questions (mapped to deployment frequency / lead time / change failure rate / MTTR + the 6 dimensions: on-call rotation maturity, IaC coverage %, observability tooling, secrets management, disaster recovery testing cadence, security scanning in CI), (b) the auto-scoring formula (each Q is 1-4, total mapped to low / medium / high / elite DORA performer band per the Accelerate report), (c) the 'where you score' results page (their band + the 3 highest-leverage fixes specific to that band), (d) the closing CTA ('want a 60-min walkthrough? book here'). Tone: senior SRE. NEVER 'unlock potential' or generic consultant filler. Output as Notion-ready Markdown."

Drop into Notion. Gate behind email. 2-3 paid discovery audits per month from this single page — at $2K/audit + $4K-$7K conversion to pipeline build-out, the math compounds fast. Worth your first quarter of revenue.

Product / Service Offering

You are selling one of three things. Pick which one before your first sales call.

DevOps audit — Two-week paid engagement, $1,500-$3,000. Read the client's CI config, IaC repo, cloud bill, and on-call setup. Score them against the four DORA metrics. Deliver a 10-page document and a 60-minute walkthrough. This is your wedge.
Pipeline + IaC build-out — Three to six weeks, $4,000-$10,000. Standardize their GitHub Actions or GitLab CI workflows, lay down Terraform or Pulumi for their AWS/GCP footprint, set up staging and production environments that mirror cleanly. Bread-and-butter.
Fractional DevOps retainer — $2,000-$5,000/month. Two days a week of on-call coverage, incident response, infra changes, and weekly office hours with their engineering team. The premium offer once you've shipped 3+ audits.

Standard stack the client will accept without argument: GitHub Actions for CI/CD, Terraform for IaC (infrastructure as code — config files that define cloud resources), AWS or GCP for cloud, Datadog or Grafana Cloud for observability, PagerDuty for on-call. Don't fight tool wars — use what the client already has and replace only what's actively breaking.

Build a reusable starter kit: Terraform modules for VPC, ECS/EKS, RDS, and S3; a GitHub Actions workflow template with build, test, lint, and deploy stages; a runbook template. By audit four, you're cloning a template and swapping variable files, not writing from scratch.

Revenue Model

Unit economics for a solo consultant running off a laptop, no employees, billing through Stripe (2.9% + $0.30 per US transaction):

Service	Price	Variable cost (cloud sandbox + Stripe)	Time	Take-home
Paid discovery audit	$2,000	$30 (AWS sandbox) + $58.30	2 weeks	~$1,910
Pipeline + IaC build-out	$7,000	$80 (sandbox + tools) + $203.30	3-5 weeks	~$6,715
Fractional retainer	$3,500/mo	$40 (shared tooling) + $101.80	2 days/wk	~$3,360/mo
Incident-response add-on	$400/hr	$11.90 Stripe per hr	as-needed	~$388/hr

Your first $1K month = one paid discovery audit at $2,000 with 50% deposit upfront = $1,000 in the door. Roughly one focused week.

Your first $3K month = one audit closed ($2,000) plus first month of a fractional retainer ($3,500, 50% deposit = $1,750) = $3,750. Roughly two weeks of work.

The retainer is where this business actually pays. Audit and build-out revenue is lumpy. Three retainers at $3,500/month is $10,500 base before any new project lands. Ten retainers is a $35K/month practice — the realistic ceiling for a solo consultant before you either subcontract or specialize harder.

Startup Costs

Cert refresh. AWS Certified DevOps Engineer – Professional is $300, with either SysOps Admin Associate or Developer Associate as a $150 prerequisite (AWS DevOps Engineer Pro). HashiCorp Terraform Associate is $70.50 and is the fastest-growing IaC credential — HashiCorp was acquired by IBM in 2024 but the cert program stays active (HashiCorp infrastructure-automation certifications). Budget $400-$650 if you need both.
Cloud sandbox + tooling. $50-$100/month across AWS Free Tier overflow, a GitHub Team seat at $4/user/month (GitHub pricing), and a Datadog or Grafana Cloud free tier for demo dashboards.
CI/CD test budget. GitHub Actions free tier covers 2,000 minutes/month for private repos; beyond that it's $0.008/minute on the standard runner (GitHub Actions billing). Knowing this in a scoping call is a credibility moment.
LLC + EIN. $35-$500 LLC filing fee by state (LLC University 50-state table). EIN (Employer Identification Number — the federal tax ID) is free at IRS EIN Online. Never pay a third party for an EIN.
E&O insurance. $800-$2,000/year for solo tech consultants via Hiscox or Insureon.
Contracts. $200-$500 attorney review of an MSA + SOW (Master Services Agreement + Statement of Work) template from Bonsai. Do this before the first signed engagement, not after.
Marketing. $0-$500. A clean LinkedIn and a one-page Notion case-study site is enough for the first six months.

Realistic all-in: $2,000 if you skip the AWS Pro cert this year, defer insurance 30 days, and lean on free tiers; $10,000 if you pay for both certs, bind E&O year one, sit a full MSA review, and prepay paid Datadog or PagerDuty seats for demos.

Legal & Formation

Business entity. Single-member LLC the moment you take a paying client. A bad Terraform apply that nukes a client's production database is a much smaller problem when the lawsuit names your LLC, not your house. Filing fee runs $35-$500 by state (LLC University 50-state table). EIN is free at IRS EIN Online — never pay a third party. Once net profit clears $80K-$100K/year, run the math on an S-corp election via IRS Form 2553. At retainer-heavy revenue this matters fast.

Licenses & sales tax. No state professional license is required for DevOps consulting. Custom consulting is generally not taxable in most states — but the moment you bundle in a hosted dashboard or recurring software access you bill for, you're closer to taxable SaaS. About 25 states tax SaaS subscriptions (Avalara state-by-state SaaS tax tracker). Cross $100K in sales or 200 transactions in any state and you have economic nexus there — the post-Wayfair threshold applies whether or not you set foot in that state. Use Stripe Tax if you bill recurring across states.

Industry-specific risk. Three things will bite this specific model. First, IP assignment and pre-existing IP carve-outs. Software you write for a client is not automatically theirs under 17 USC §101 — you need an explicit IP assignment clause. But your reusable Terraform modules, runbook templates, and GitHub Actions workflows are your pre-existing IP. Call them out in your MSA as retained by you, and grant the client a license to use them inside their environment. Without that carve-out, every engagement hands away tooling you spent months building. Second, cap on liability for production outages. A misconfigured deploy can cost a client real revenue — an e-commerce client losing a day of sales is a six-figure number. Cap your total liability at fees paid in the prior 12 months, in writing, in the MSA. Never accept uncapped indemnification. Third, scope creep on Kubernetes engagements. A "move us to Kubernetes" engagement routinely expands 3x when the client's app architecture isn't containerization-ready. Don't promise a production K8s migration as a solo project under $5K — the realistic floor is $8K-$20K and several months. Always start with a paid $1,500-$2,000 discovery audit before scoping anything larger. The audit is the gate.

Marketing & First Customers

Your first 3 paying clients come from people who already know you can ship. After that, four channels work for this niche.

Direct outreach to 30 Series A SaaS startups, filtered tight. Use Crunchbase or Apollo (~$49/month) to filter to SaaS companies that closed a $3M-$15M Series A in the last 9 months with 8-25 engineers on LinkedIn. Send a 4-line LinkedIn message to the CTO: name a specific DORA metric you'd help them improve, attach a 2-page sample audit, offer a 20-minute call. Target: 10-15 messages/week. Reply rate 8-12%, call-to-paid 15-25%.
Hacker News "Ask HN: Who wants to be hired?" and "Who is hiring?" threads. Post monthly on the first business day. A short paragraph naming your stack, your DORA-baseline-then-fix engagement, and a price range pulls 2-4 inbound leads per thread. Free, high signal (Hacker News Show HN).
One detailed public case study per quarter. Write up a real audit-and-fix — deployment frequency before and after, the specific GitHub Actions and Terraform changes — and post to dev.to and LinkedIn. Anonymize the client. These rank for "how to fix slow deploys" long-tail searches and pull warm inbound for 18+ months.
PagerDuty and Datadog partner programs. Both have referral tiers for consultants who implement their tools at client sites (PagerDuty Partners, Datadog Partner Network). After 3-5 implementations, applying gets you co-marketing and warm inbound from their sales team.

Disclose your AI tool stack — Copilot, Cursor, Claude Code — upfront in scoping calls. Saying it first builds trust and surfaces objections before contract. GitHub Copilot for Business includes commercial-use IP indemnification.

First 90 Days

Week 1. File LLC. Get free EIN. Pick AWS or GCP as your primary cloud and commit. Do not learn both at once.
Week 1-2. Refresh or earn the AWS DevOps Engineer Pro ($300) or HashiCorp Terraform Associate ($70.50). One credential on the LinkedIn profile is enough at this stage.
Week 2-3. Build your reusable Terraform module library and a GitHub Actions workflow template. Push to a public starter repo with a clean README. This is your portfolio.
Week 3-4. Attorney reviews MSA + SOW. Bind E&O insurance via Hiscox. Write your 2-page audit deliverable template and your 1-page scope-control checklist.
Week 4-5. Build a free 10-question DORA self-assessment as a Notion page or Typeform. Use it as a lead magnet on LinkedIn and in cold outreach.
Week 5-7. Send 30 cold messages to filtered Series A SaaS CTOs. Land 1 paid discovery audit at $1,500-$2,500 in exchange for a written case-study testimonial.
Week 7-10. Ship the audit. Capture a 10-minute Loom walking through the DORA scoring and the recommended fix list. Post a sanitized version to LinkedIn.
Week 10-12. Convert the audit into a $4K-$7K pipeline build-out OR a $2K-$3K/month fractional retainer. Land 1 more audit. Target: 2-3 paying clients, 1 retainer signed, ~$5,000-$9,000 in 90-day revenue.

Common Pitfalls

Quoting hourly without a cert and getting bracketed at $80/hour. Rates above $200/hour assume 3+ years of senior DevOps work and an AWS or GCP professional-tier credential — without one, clients anchor you to junior contractor rates. Spend the $300 on the AWS DevOps Pro exam before your first proposal goes out. Less than one billable day to add $50/hour to every quote for two years.
Promising a Kubernetes migration as a fixed-price under $5K. Realistic floor for a production K8s migration is $8K-$20K and several months — and it routinely expands 3x when the client's app isn't containerization-ready. Always run a paid $1,500-$2,000 discovery audit first. Quote the migration only after the audit confirms feasibility, with a 25% scope-buffer line. Skip this and you'll burn 6 weeks of unpaid overruns on the first big engagement.
Skipping the IP carve-out on your reusable modules. Without an explicit "pre-existing IP" clause in the MSA, a client could argue your Terraform module library belongs to them once they paid for an engagement that used it. Bake a 3-paragraph carve-out into your master MSA and have the attorney review it once. The $200-$500 review is cheaper than rebuilding your toolkit because a client's lawyer asked.
Letting client cloud spend run through your AWS account. When their traffic spikes, your $40 monthly sandbox bill becomes $4,000 on your card and you wait 60 days for reimbursement. Either bill the client's account directly via AWS Organizations as a delegate, or pass cloud costs through with a 15-20% management margin and a monthly cap in the SOW. Decide before signing — switching mid-engagement ends with someone eating an unpaid AWS bill.

Get your full launch plan — take the free 60-second quiz.