QA & Playwright Class Notes by Batch

AI Tester Blueprint 4x

12 classes

The 4x batch and the AI Tester group run together, so their classes are listed as one track.

Rebuilding the naive RAG in Langflow, choosing a vector database, and the MCP certification plan

AI Tester Blueprint 4x notes: the naive RAG rebuilt in Langflow, local versus cloud hosting, vector database and free embedding options, the component flow, and the MCP plan.

Sat 26 Sep · 2h 0m

Building a naive RAG in n8n and Pinecone, and the chunking bug that split test cases in half

AI Tester Blueprint 4x notes: a naive RAG over 100 test cases in n8n and Pinecone, embeddings, chunk size and overlap, and fixing a chunking bug with a CSV cleaning step.

Sat 19 Sep · 1h 48m

What RAG is, the two phases it runs in, and the RAG Explorer

AI Tester Blueprint 4x notes on RAG for QA: the three limits of a language model, fine-tuning versus RAG, the ingestion phase, the retrieval augmentation and generation phase, vector databases and embeddings, and a Langflow RAG pipeline over 500 test cases.

Sat 12 Sep · 2h 1m

A bug triage agent on a canvas, then the same flow behind curl, Postman and a React page

AI Tester Blueprint 4x notes on Langflow: build a bug triage agent, fetch a Jira issue over REST with base64 basic auth, parse the response into a prompt template, make the ticket key dynamic, and call the finished flow from curl, Postman, Claude Code or a React UI.

Sun 6 Sep · 1h 51m

A screenshot goes in, a Jira bug comes out

Build a screenshot to bug reporter in n8n: form trigger, base64 conversion, a Groq vision model, a Jira ticket with the screenshot attached. Debug rate limits, a bad model id and an expired Jira token live, publish a production webhook, front it with a Vercel UI through a serverless proxy, then install Langflow and build a first flow.

Sat 5 Sep · 2h 3m

Building an agent live, and the four times it broke

Building an n8n content agent end to end: the trigger, action, result framing, structured JSON output from an AI agent node, Pin Data to save quota, the Gemini image API through an HTTP Request node, and a Telegram human-in-the-loop approval before anything is published.

Sun 30 Aug · 1h 55m

Running n8n on your own machine, and the six agents built on it

Install n8n locally with npm and run it against Ollama for a free setup, then build six QA agents: Jira fetch and create, a local test case generator, bug triage to Google Sheets, RCA, and a social media poster.

Sat 29 Aug · 2h

BLAST, and the first agent that runs without you

The BLAST framework for building trackable AI agents, phase by phase, plus an n8n introduction: TAR, the two Jira agents built live, choosing a model, what it costs, and the rule about company laptops.

Sun 23 Aug · 1h 52m

Branding for testers: the profile, and the content engine behind it

LinkedIn branding for QA engineers: how to edit your profile without alerting your employer, what recruiters actually search, the Hook-Story-Offer post structure, building a content brain and a reusable brand voice, and a 60-day posting challenge.

Sat 22 Aug · 1h 53m

One resume per job: beating the ATS with AI

Tailor your QA resume to each job description with AI: how applicant tracking systems filter, the five things they score, the resume structure that works, and a repeatable skill that turns a job description into a matched resume.

Sun 9 Aug · 2h 07m

Local LLMs for QA, and building a Jira test case generator

Run local LLMs for QA with Ollama and LM Studio: open-source vs cloud models, and vibe-coding a Jira-connected test case generator with a Groq fallback.

Sat 8 Aug · 2h 21m

Prompt engineering for QA: the RICEPOT framework

Prompt engineering for QA testers: the RICEPOT framework, CRAIP, RACE and COAST alternatives, zero-shot vs few-shot, and ready-to-use QA prompt templates.

AI Tester Blueprint 3x

10 classes

The 3x batch track: Python fundamentals through to the AI frameworks.

Sat 19 Sep · 54m

The end-to-end QA co-pilot, and what an agentic run costs

AI Tester Blueprint 3x notes: a LangChain QA co-pilot that fetches a Jira ticket, retrieves similar test cases from a local RAG, plans, executes in Chromium and reports to Slack, plus cost per agentic run and when to choose LangChain over Crew.ai.

Sat 12 Sep · 1h 47m

Bricks, a home and the neighbourhood aunty, then an agent that drives a real browser

AI Tester Blueprint 3x notes on LangChain with Python: LangChain versus LangGraph versus LangSmith, create_agent with Groq and Gemini, system prompts, parallel calls with asyncio.gather, custom tools, Pydantic structured output, and a Playwright orchestrator agent.

Sun 6 Sep · 2h 11m

From one passing test to a framework that scores 25 metrics

Build a DeepEval framework end to end: golden datasets for a chatbot and a RAG, an independent judge model on Groq, 25 metric specifications covering quality, safety, retrieval and security, a dashboard with token accounting, and the settled answer on which direction a hallucination score points.

Sat 5 Sep · 1h 35m

Thresholds, and why every DeepEval score points the same way

LLM evaluation with DeepEval: why evaluation is needed at all, thresholds as acceptance criteria, why every metric scores the good direction so the threshold is always a floor, mapping DeepEval vocabulary onto ordinary QA terms, environment setup, using Groq as the judge model, and a first answer relevancy test case.

Sun 30 Aug · 1h 22m

AI is not tested, it is evaluated

Why LLM systems are evaluated rather than tested, the vocabulary of LLM evaluation, DeepEval versus Ragas, thresholds and LLM-as-judge, and where red teaming fits.

Sat 29 Aug · 1h 30m

Three agents for the meeting nobody wants to attend

Building a CrewAI bug triage crew for QA: three agents for classification, root cause and test recommendations, Jira as a tool, sequential process, provider fallback, and the one agent one task rule.

Sun 23 Aug · 1h 37m

Finishing Python, and building your first CrewAI agent

Finish Python and build your first CrewAI agent, with the full pytest and CrewAI reference on the same page: fixtures, parametrize, markers and the flags worth knowing, agent, task and crew fields, sequential versus hierarchical, the model-provider table, and how to actually connect Groq.

Sat 22 Aug · 1h 30m

Finishing Python: polymorphism, abstraction, statics and exceptions

Python OOP for AI and QA work: why Python has no method overloading, how overriding differs, abstraction with ABC and abstractmethod, static versus instance attributes, and try/except/else/finally including the else clause Java does not have.

Sun 9 Aug · 1h 26m

Python OOP: constructors, encapsulation, and the five kinds of inheritance

Python OOP class notes for testers: __init__ constructors, self, encapsulation with public, protected and private, five inheritance types, and MRO explained.

Sat 8 Aug · 1h 36m

Python sets, dictionaries, map and filter, and the start of OOP

Python sets and dictionaries for SDET interviews: set operations, map vs filter, dict.get counting tricks, zip, and the start of classes and objects.

Playwright 4x

3 classes

The Playwright 4x batch: ninety days of JavaScript, TypeScript, Playwright and AI, from the 28 September kickoff.

Fri 2 Oct · 1h 23m

Prompt engineering for QA: five principles, zero-shot and few-shot prompts, the RICE-POT framework, and a first push to GitHub

Playwright 4x notes: prompt engineering for QA, five principles, zero-shot and few-shot prompting, the RICE-POT framework and template, and pushing to GitHub.

Wed 30 Sep · 1h 22m

AI engineering, how LLMs use tokens and parameters, the core AI glossary, and anti-hallucination rules

Playwright 4x notes: AI engineering vs machine learning, LLM tokens and parameters, Ollama, the core AI glossary, RAG, and anti-hallucination rules for QA.

Mon 28 Sep · 1h 9m

Batch kickoff: the 90-day roadmap, where everything lives, and the day-one setup tasks

Playwright 4x kickoff notes: the 90-day roadmap, the progress dashboard, SDET Club, notes and recordings, the three repositories, the AI tool policy and the day-one tasks.

Playwright 3x

20 classes

The Playwright 3x batch, running on its own weekday schedule.

Thu 1 Oct · 1h 8m

Keyboard presses, hover, drag and drop, right-click menus, and JavaScript alerts with page.once() and page.on()

Playwright 3x notes: keyboard presses, hover, dragTo and a boundingBox mouse drag, right-click menus, and JavaScript alerts with page.once and page.on.

Tue 29 Sep · 1h 6m

Native, custom and async dropdowns, and iframes, framesets and nested frames with frameLocator()

Playwright 3x notes: selectOption for native dropdowns, getByRole for custom and async menus, and frameLocator for iframes, framesets and nested frames.

Sat 26 Sep · 1h 0m

locator.filter(), the :has() pseudo-class, and searching a paginated web table

Playwright 3x notes: locator.filter() with hasText and has, the CSS :has() pseudo-class for table rows, searching a paginated table, and a reusable findRowByName.

Thu 24 Sep · 1h 10m

Multiple elements, the three text getters, and web tables with dynamic XPath

Playwright 3x class notes: all versus allInnerTexts versus allTextContents, strict mode and first, web table anatomy, dynamic XPath with nested loops, and following-sibling.

Sat 19 Sep · 1h 6m

CSS selectors, the three run modes, and getByRole

Playwright 3x class notes: the CSS selector cheat sheet, UI mode versus debug mode versus headless, the locator priority ladder, ARIA roles, and getByRole with the accessible name, including why name is not the HTML name attribute and what exact true changes.

Thu 17 Sep · 1h 8m

XPath, strict mode, and locators for elements that change

Playwright 3x class notes on XPath: absolute versus relative, the locator confidence ranking, contains starts-with and normalize-space, handling dynamic ids, XPath axes for web tables, and Playwright strict mode with first, last and nth.

Sat 12 Sep · 1h 3m

Five annotations, one grouping keyword, and the start of finding elements

Playwright 3x class notes: the full course roadmap, browser context page recap with three roles, browser context options including mobile emulation, test.skip, only, fail, fixme and slow, test.describe grouping with the -g flag, and an introduction to the two locator strategies.

Thu 10 Sep · 1h 18m

How a Playwright command reaches the browser, and why that beats one HTTP request per step

Playwright 3x class notes on architecture: Selenium's stateless W3C requests vs Playwright's persistent WebSocket, JSON-RPC, CDP vs Playwright protocol (CDP+), why Chromium is the default, channel and Chrome for Testing, and browser, context and page.

Tue 8 Sep · 1h 10m

Why Playwright, what npm init gives you, and a test you recorded instead of wrote

Playwright 3x kickoff notes: why teams moved from Selenium, rendering engines versus real browsers, Playwright's real limitations, npm init playwright@latest and what it creates, headless versus headed, the Playwright CLI, and recording a first test with Codegen.

Sat 5 Sep · 1h 1m

Enums, generics and access modifiers, and which of the three survives compilation

TypeScript enums, generics and access modifiers for test automation: named constants for statuses and environments, generic functions and classes, why T beats any, public against private against protected in a page object model, readonly, and abstract classes.

Wed 3 Sep · 1h 19m

Abstraction, interfaces, and the readonly that does not actually lock anything

TypeScript abstraction and interfaces for test automation: how abstraction differs from encapsulation, interfaces as contracts, readonly and optional fields, methods in interfaces, interface extends interface against class implements interface, and why interfaces vanish at compile time.

Tue 1 Sep · 1h 6m

Polymorphism, and the TypeScript setup that breaks on a fresh install

JavaScript inheritance types and polymorphism for test automation, then the start of TypeScript: which inheritance shapes JavaScript allows, mixins, method overriding and dispatch, type annotations, void, never, any and unknown, and how to actually run a .ts file.

Sat 29 Aug · 1h 3m

Encapsulation, inheritance, and the super that quietly returns undefined

JavaScript encapsulation and inheritance for test automation: private fields with getters and setters, controlled writes, static private counters, the five inheritance shapes, super for constructors and methods, and method overriding.

Thu 27 Aug · 1h 6m

Constructors, private fields, and the one keyword to avoid in Playwright

JavaScript constructors, this, private hash fields and static explained for test automation, with the verified reason to avoid static in Playwright page objects: static state lives per worker process, so counters silently under-report.

Tue 25 Aug · 1h 15m

Export, import, and the class that replaces fifty exports

ES modules and classes for test automation: named versus default exports, alias imports, why import star is a bad habit, and why a class beats fifty exports. Plus what new actually does, private hash fields, object references, and the move from JavaScript to TypeScript.

Sat 22 Aug · 1h 5m

Two rules for async/await, and what happens when you forget one

Async and await in Playwright with two mechanical rules, plus the event loop that explains flaky tests, why promises beat setTimeout, try/catch/finally, a retry loop, and three outdated habits including the deprecated waitForNavigation.

Thu 13 Aug · 1h 32m

Callbacks, promises, and async/await: why Playwright needs all three

Callbacks, promises and async/await for Playwright testers: callback hell, the three promise states, then, catch, finally, Promise.all, allSettled, race, and why Playwright is asynchronous.

Tue 11 Aug · 1h 23m

JavaScript 2D arrays and the star patterns interviews love

JavaScript 2D arrays and star pattern programs for SDET interviews: nested loops, process.stdout.write, triangles and the pyramid with full trace tables.

Sat 8 Aug · 1h 03m

JavaScript strings, objects, and call by reference

JavaScript string methods for SDET interviews: slice vs substring, immutability, reversing a string with a loop, objects vs JSON, and call by value vs reference.

Thu 6 Aug · 1h 26m

Closures: the variable that outlives its function

JavaScript closures explained for automation testers: hoisting and TDZ recap, function versus block scope, and three closure examples (makeCounter, retry limiter, rate limiter) with the exact output each one prints.

Playwright 2x

13 classes

The Playwright 2x batch, running on its own weekday schedule.

Mon 14 Sep · 1h 12m

A gatekeeper for a framework everyone is now writing with AI

Playwright 2x class notes on quality gates: ESLint 9 flat config with type-aware rules for a Playwright suite, lint and typecheck scripts, and four agent-readable gates (AI slop, ponytail, over-engineering, framework patterns) that block a pull request built by Copilot, Claude or Cursor.

Fri 11 Sep · 1h 8m

An AI layer for the framework: one LLM client, one factory, four agents, no test that trusts the model

Playwright 2x class notes: add an AI layer to a Playwright framework with one LLM client for DeepSeek, OpenRouter, Groq, OpenAI and Claude, an agent factory with schema validation and one retry, an AI data generator, RCA and flaky agents in the reporter, and self-healing locators.

Wed 9 Sep · 1h 13m

The shape check that runs before every assertion, and how many workers your laptop can really take

Add JSON schema validation to a Playwright API framework with Ajv: what a schema checks, the strict create-booking schema, a shared validator, the API test order (ping, schema, CRUD, end to end), trimming AI-generated code with Ponytail, and a measured guide to Playwright workers, RAM and sharding.

Mon 7 Sep · 1h 17m

Three layers, and a spec that is only assertions

Refactor a Playwright API suite into three layers: an ApiHelper that builds URLs and calls the API, a BookingApi service object and a fixture that supplies the token, and specs that only assert. Includes managed token renewal on 403, negative tests, and a JSONPath cheat sheet with jsonpath-plus.

Fri 4 Sep · 1h 2m

CRUD end to end, and why the failure is 403 and not 401

Automating a full CRUD flow in Playwright: auth token, create booking, update booking, with typed interfaces, test.step and shared flow state, the Cookie header for authorization, request.newContext for a second base URL, and the arrange-act-assert refactor that follows.

Wed 2 Sep · 1h 9m

API testing starts here, and the status codes that will fail your first assertion

Playwright API testing for the 2x batch: the three-layer architecture, URL anatomy, HTTP methods and status families, why Playwright beats Postman and Rest Assured for automation, the request fixture and APIRequestContext, and the config change that makes API tests run.

Fri 28 Aug · 1h 00m

Adding one feature with AI, without becoming a blind AI user

Add dotenv support to a Playwright framework using an AI coding agent properly: plan mode first, review and push back on the plan, check the blast radius before touching shared files, catch the gitignored .env that breaks CI, approve file by file, then prove the values are really being injected.

Wed 26 Aug · 1h 13m

Fixtures: handing every test a page that is already logged in

Playwright custom fixtures explained: fixtures versus hooks, the built-in page, context, browser and request fixtures, building test-base.ts with test.extend, what use() does, environment-variable credentials, a visualStep wrapper for per-step screenshots, and the trace viewer.

Mon 24 Aug · 1h 15m

The build plan: how to rebuild the whole framework from an empty folder

The seven-phase plan for rebuilding an advanced Playwright TypeScript framework from an empty folder: project structure, config and environments, utilities, page objects, specs, reporting and quality gates, CI and AI rules, plus the interview answers from the class and a verified audit of what the repo contains today.

Wed 12 Aug · 1h 33m