Reserve your Spot →
New Course  ·  Cohort 1 starts Sep 1, 2026

Lead the AI transition in your QA team - without being replaced by it

A 5-week advanced course for working QA Automation engineers. Learn to govern Claude Code as a reproducible AI teammate inside a Playwright + Jira + Zephyr workflow - and become the most valuable engineer in the room.

Reserve your Spot → See 5-week plan
CLAUDE.md  ·  the AI constitution
# Workflow Constitution # Read this on every session. ON task_type IS "new_test": read(10-playwright-pom.md)
read(11-fixtures.md)
ON task_type IS "zephyr_writeback": read(51-zephyr-squad.md)
REQUIRE human_approval == "true" ON step_complete: update(vault/wiki/) # non-skippable
✓
Bounded behavior Auditable, gated outcomes - not randomness
5 wks
Duration
~22 hrs
Video content
~5 hrs/wk
Working pro pace
Mid+
For QA / SDET
The course in 60 seconds
NOT ANOTHER PLAYWRIGHT TUTORIAL

In 2026, mid-level QA engineers aren't paid to write tests by hand anymore. They're paid to design, govern, and audit a system where AI writes most of the tests under their supervision. That's exactly what this course teaches.

❌ What we don't teach
  • How to install Playwright for the first time
  • What a Page Object Model is conceptually
  • Generic "AI for testing" theory
  • Tool tutorials you can find free on YouTube
  • "Just paste this prompt and AI does everything"
✅ What you actually learn
  • A versioned instruction-file constitution that governs Claude Code - so AI works inside YOUR rules, not outside them
  • The Karpathy-style knowledge vault that compounds knowledge across every run
  • Gated writeback to Jira + Zephyr with a human approval contract - no accidental pushes
  • AI test evaluation - judging output quality, not just generating it
  • Self-healing for flaky tests in CI - local-triggered, no extra API costs
Why this matters
THE 2026 REALITY FOR QA ENGINEERS
📈
58%
of enterprises actively upskilling QA teams in AI tools right now
⚡
55%
of QA teams report a skilled-labor shortage in automation
💰
15%
salary premium for Playwright roles over Selenium in 2026 job postings
🚀
3-5×
higher ROI than ISTQB cert. US hiring managers want portfolios, not papers.
The plan
5 WEEKS. 19 MODULES. ONE WORKFLOW.

~22 hours of video plus reading and hands-on practice. Weeks are intentionally uneven - some heavy, some light. The buffer is by design, not accident.

Week 01 🧠 Foundation - mindset, setup, AI safety ~2.5h · 2 modules +
We frame what AI-augmented QA actually means in 2026, why frontier models aren't deterministic, and why bounded behavior matters more than raw automation speed. Then we get the toolchain on your machine.
  • Module 1: Mindset, landscape, session hygiene, AI safety
  • Module 2: Environment and Claude Code setup, first CLAUDE.md skeleton
Week 02 - INTENSIVE ⚙️ Raw → refactor → MCP ~4.5h · 3 modules +
The most important week. You'll write raw Playwright tests, feel the pain of duplication, refactor into a clean POM + fixtures + data architecture with Claude as pair, then meet Playwright MCP - and see it generate a draft that breaks your POM. That observation seeds the constitution next week.
  • Module 3: First two raw Playwright tests (Claude-assisted, no MCP yet)
  • Module 4: Pain-driven refactor - POM, fixtures, user data, utils
  • Module 7: Playwright MCP - once the framework exists
Week 03 📜 The instruction-file constitution + multi-env ~2.75h · 2 modules +
The differentiator. You write CLAUDE.md as a router and author the first four instruction files. You then try to break your own rules - and tighten the wording when Claude finds the loopholes.
  • Module 5: Writing the instruction files (CLAUDE.md as router)
  • Module 6: Multi-env setup - dev (full coverage) + prod (safe smoke only)
Week 04 🗄️ Karpathy vault - persistent memory + data layer ~2.75h · 2 modules +
Set up an Obsidian vault inside your repo using the Karpathy LLM-Wiki pattern. Then wire local Postgres in Docker for real test data seeding and teardown that doesn't bleed between tests.
  • Module 8: Karpathy vault and obsidian-vault.md instruction file
  • Module 9: PostgreSQL MCP + data management rules
Week 05 - INTENSIVE 🚀 CI, Jira + Zephyr, capstone - ship it all ~7h · 11 modules +
The final sprint. API tests from Swagger, Docker parity, Allure reporting, the Jira + Zephyr closed loop with gated writeback, GitHub Actions CI with self-healing, visual + a11y, and your capstone. You leave with a portfolio employers actively search for.
  • Module 10: API testing with Playwright APIRequestContext + Swagger
  • Module 11: Suites + Docker parity
  • Module 12: Allure reporting + 2-week history retention
  • Module 13: Jira + Zephyr Squad integration + writeback gate
  • Module 14: GitHub Actions CI + local-triggered self-healing
  • Module 15: Visual regression + accessibility scans
  • Module 16: AI-generated test evaluation (5-axis rubric)
  • Module 17: Jira dashboards + coverage gadgets
  • Module 18: Full workflow walkthrough + unhappy-path branch
  • Module 19: Capstone + what's next
What you walk away with
13 PORTFOLIO ARTIFACTS MOST CANDIDATES CAN'T SHOW

Every competing course graduates students with a generic Playwright tutorial repo. You leave with things that are individually rare and collectively unique - including the instruction-file constitution that hiring managers are starting to ask about by name.

01
⌨️ Complete Playwright + TS frameworkPOM, fixtures, data, utils - built by you, not handed to you
02
📜 CLAUDE.md + 12+ instruction filesThe constitution that governs every AI action in your workflow
03
🧠 Karpathy-style knowledge vaultSelf-maintaining Obsidian wiki with pages for every test and debug session
04
🌐 Multi-env configurationDev (full coverage) + prod (safe smoke) - with hard guardrails
05
🗄️ Postgres MCP wiringTest data seeding and teardown via Model Context Protocol
06
🔗 API test suite from SwaggerGenerated, evaluated, and tightened under the constitution
07
🐳 Suite tagging + Docker paritySmoke / regression / api / e2e - runs the same on every machine
08
📊 Allure reporting with history2-week trend graphs that survive CI churn
09
🔄 Jira/Zephyr closed loopGated writeback workflow with full audit trail
10
⚙️ GitHub Actions CI + self-healingDeterministic execution plus local-triggered maintenance
11
👁️ Visual regression + a11y suitePlaywright screenshots + axe-core coverage
12
📋 Personal Jira dashboardThe kind your manager will actually screenshot for standup
13
🏆 Capstone repositoryYour own user-story-to-Done walkthrough, ready for any interview
Why now
THE NARROW, WELL-PAID SPACE - AND IT'S WIDENING FAST

AI is squeezing salaries for engineers who only write tests. It's widening salaries for engineers who own the workflow design. This course targets exactly that gap.

$115K

Mid QA / SDET - US median (2026)

Mid-level QA Automation engineers earn $95K-$135K. Playwright stack pulls 10-15% above Selenium equivalents in current postings.

$155K+

Senior SDET / Automation Architect

Engineers with AI workflow fluency command $130K-$185K+ in 2026. Top-of-band with a governance portfolio see a 20-25% premium.

3-5×

Higher ROI than ISTQB

US hiring managers rarely ask for certs. They ask what you've built - and how you'd govern an AI doing it. A portfolio wins every time.

2026

The transition year - act before the window closes

The role is shifting: "QA writes tests" → "QA governs AI that writes tests." Supply of qualified engineers is still very small. Being early is the moat.

The engineer who occupies the space between "lets AI run unchecked" and "refuses to use AI at all" is the engineer who keeps their job - and gets the senior offer. That space is exactly what we teach.
Be honest with yourself
WHO THIS IS FOR - AND WHO IT ISN'T
✅ This is for you
  • ⌨️
    You've written automated tests in any framework - Playwright, Cypress, Selenium, WebdriverIO, Pytest
  • 🔧
    You're comfortable with JS/TS basics, Git, the terminal - no need to be a wizard
  • 🤖
    You see AI changing your job and want to be the engineer who shapes it - not the one it replaces
  • 💼
    You want a portfolio artifact employers are actively hiring for right now in 2026
  • ⏱️
    You can commit ~5 hours per week for 5 weeks - no nights and weekends, just focused time
⛔ Not for you (yet)
  • 🚫
    You've never written an automated test. Start with Codemify's Manual QA → Automation QA path first
  • 📺
    You want a step-by-step Playwright tutorial. There are great free ones - this isn't that
  • 🎰
    You want "paste this prompt and the AI does everything." Frontier models aren't deterministic; we teach you to constrain them
  • ⏳
    You can't commit time for 5 weeks. Self-paced is available, but cohort learning compounds 3x faster
⚡ AI-powered workflow

Your AI teammate is waiting.
You just need to give it rules.

Claude Code won't replace you. But an engineer who knows how to govern Claude Code will replace the engineer who doesn't.

5 wks
Duration
13
Portfolio artifacts
$1,250
Early bird
Sep 1
Cohort starts
Reserve your Spot →
The stack

REAL TOOLS. 2026 JOB POSTINGS.

Every tool in this course appears in current QA/SDET job descriptions. No toy projects, no made-up stacks.

⌨️
Playwright + TS
Browser automation
🤖
Claude Code
AI teammate · Pro plan
🗄️
PostgreSQL MCP
Test data seeding
🔗
Playwright MCP
Guided exploration
📋
Jira + Zephyr
Test management
🐙
GitHub Actions
CI/CD pipeline
📊
Allure
Reporting + history
🧠
Obsidian vault
Karpathy LLM-Wiki
Early bird pricing

CHOOSE YOUR PACE - SAME CONTENT, DIFFERENT INTENSITY

Both tracks cover the full 5-week curriculum and give you all 13 portfolio artifacts. The difference is how much support you get along the way.

Video lessons only

Self-paced - learn on your schedule

All the videos, instruction-file templates, and sandbox access - on your timeline, wherever you are.

$500 $250
🎯 Early bird - first 50 people only
  • ~22 hours of video, lifetime content access
  • All instruction-file templates and skeleton starters
  • Sandbox access for 8 weeks
  • Codemify community Discord
  • Self-graded capstone rubric
Instant access after purchase
Recommended - live mentorship

With mentors - real feedback, real results

Live group sessions, 1:1 mentor time, capstone review, and career coaching from Sergii's team.

$2,500 $1,250
🔥 Early bird - first 20 people only
  • Everything in self-paced
  • Weekly live group session (~90 min)
  • 30-min 1:1 with your mentor every 2 weeks
  • Mentor review of your capstone
  • Resume + interview prep
  • Cohort 1 alumni discount on future courses
Cohort 1 starts Sep 1, 2026 · 100% refund within 3 days
Honest answers

QUESTIONS YOU SHOULD BE ASKING

Do I really need prior automation experience?
Strong familiarity is enough. If you've written automated tests in any framework - Playwright, Cypress, Selenium, Pytest - you'll keep up. Complete beginners can enroll but should plan for extra self-study on Playwright syntax and TypeScript. We move quickly through fundamentals because the course is the AI workflow on top, not the fundamentals themselves.
Do I need an Anthropic API key?
No. The entire course runs on a Claude Pro subscription (~$20/month) or a Team plan if your employer provides one. We deliberately designed every workflow to fit inside subscription limits. Module 1 teaches session hygiene so you stay comfortably under the weekly cap.
Will Claude Code give every student the same output?
No - and you shouldn't want it to. Frontier models aren't deterministic. What we teach is bounded behavior: a versioned instruction-file constitution that constrains outcomes (what gets shipped, what gets gated, what gets written back) rather than paths (what Claude types next). Same mental model a senior engineer uses to manage a junior: clear contract, hard gates, generous freedom in the middle.
Is the Jira/Zephyr sandbox provided?
Yes. Codemify provisions per-student Jira projects inside a shared Zephyr Squad instance for the duration of the course. You don't need to bring your employer's sandbox or pay Atlassian separately.
How is this different from Codemify's existing Automation QA course?
The Automation QA course teaches you to become a QA Automation engineer from scratch - Playwright, frameworks, CI, your first job. This course assumes you've already done that (or have equivalent experience) and teaches the next layer: governing AI inside that workflow. Different audience, different price, different promise. Think of it as the senior track.
What if my company uses Cypress or Selenium?
The Playwright pieces transfer cleanly. The instruction-file architecture is framework-agnostic - the constitution governs Claude regardless of which test runner sits underneath. Many students adapt the patterns to their employer's existing stack within their first sprint back.
What's the refund policy?
100% money back within 3 calendar days of cohort start. Partial refund on a pro-rata basis days 4-14. No refunds after 14 days. The same Codemify policy that applies to all our courses.
Will the course be updated as AI tooling changes?
Yes. The curriculum is intentionally centered on workflow architecture, which moves slowly, rather than specific tool UI, which moves weekly. The non-portable bits (specific CLI flags, MCP server endpoints) get patched as needed. Lifetime content access means you receive updates without paying again.
Cohort 1 · Limited 20 seats · Early bird pricing ends soon

Lead the transition. Don't get caught by it.

5 weeks. One workflow. A portfolio most candidates can't show. Talk to Sergii and see if it's the right fit - no hard sell, just an honest 30-minute conversation.

September 1, 2026  ·  100% refund within 3 days  ·  Early bird: $1,250 (reg $2,500)