Alexander Degai

Everyone’s saying QA is dead. They’re wrong. It’s just being rebuilt by AI — quietly, powerfully, and permanently. This isn’t a debate about tools. It’s a transformation of how we think, test, and ship with confidence. Here’s how modern QA is evolving: – Tests are written by prompt – Root causes are explained, not guessed – Test data is smarter, contextual, fast – Only the tests that matter are run – AI spots bugs before the UI is even pushed – And automation feels like architecture — not repetition I’ve mapped the exact shift in this 8-slide guide👇 💬 What part of this AI evolution are you most excited about? #AIinQA #Playwright #TestAutomation #ModernTesting #SoftwareTesting #SDET #QATransformation #PromptEngineering #AutomationStrategy #EngineeringLeadership

265

29 Comments

Ok. I'll say it. With so much talk about AI and QA and AI replacing QA and the future of QA, does it seem to anyone else that there are 798,000 generic profiles posting the same regurgitated articles about "AI and QA" and "AI replacing QA" and "the future of QA"? It's almost, as if, AI is doing alot of what we're reading here on LinkedIn. I block a dozen or so "profiles" daily just so I can read something a human has written. But Jason, how can you tell? They all start the same. with a generic entry paragraph. all paragraphs are spaced. there are ZERO grammatical errors. and there's always a bulletpointed list in their posts. I certainly don't think there's any monetary value in posting to LinkedIn. There isn't a way to monetize engagement so... why? I don't know. Maybe it's just old man Jason yelling at the kids to get off my lawn but... this site it circling the drain with the amount of redundant profiles and posts. It dilutes the field.

191

52 Comments

Most QA tools stop at reporting bugs. The best ones learn from your feedback. Here’s a tactic we’ve been using at Pie to make AI testing smarter over time — without overwhelming the user: 📌 Every dismissed issue becomes training data. If you mark a finding as “low priority” or “expected behavior,” we store that context and use it in future runs. 🔁 The agent adapts. It won’t flag the same false positive again. Instead, it uses your signal to improve its precision — for your app. 🧠 Next: fine-tuning per customer. We’re training app-specific models so the more you use the system, the more tailored and context-aware it becomes — like onboarding a QA engineer who remembers your Apps quirks and edge cases. 💡 Pro tip: Want faster improvement? Add a short reason when you dismiss an issue. It makes the learning loop tighter and more accurate. Most QA automation tools are read-only. The best ones are collaborative. #qaautomation #aiagents #productops #devtools #buildsmarter #b2bsaas

40

1 Comment

If you are vibe coding, you need to 2x or 3x your QA staff. Why? Because you are pumping out more code, more features, and getting them ready for review much faster. This coding is also lower quality and can cause a lot more issues. This means that there are going to be a lot more QA issues, more tickets, and more testing. So make sure you are budgeting for extra QA if you are expecting your engineers to vibe code.

507

328 Comments

Think QA is just about squashing bugs?   Think again.   There's a bigger picture... Quality assurance isn't just bug-hunting. It's so much more. It's about alignment. Understanding how things affect the business, engaging the right people, and making sure testing meets user needs. It's not just about finding problems. It's about stopping them before they even start. Bugs? Just symptoms. The real issue? Poor quality strategy. Want to avoid common QA pitfalls? Here’s what to watch for: - Lack of clear communication - Ignoring user feedback - Rushing the testing phase - Not involving stakeholders - Overlooking long-term impact Quality can't be an afterthought. So, what's the biggest QA mistake you see every day? Let's talk!  #QualityAssurance #SoftwareTesting #TechIndustry

9

Evolving our AI-powered QA workflow One of the biggest frustrations in test automation is flaky tests. They break when locators shift, environments change, or data drifts even though the application itself is working fine. That’s why we’ve been building self-healing tests directly into our AI QA workflow. Here’s how it works: 1️⃣ Requirements go in → test cases are generated with LLMs. 2️⃣ Playwright automation is created at the click of a button. 3️⃣ Tests run and feed results back into the workflow. 4️⃣ When a test breaks, the AI doesn’t just fail. It attempts to heal itself — automatically updating selectors, adjusting to UI changes, and re-validating. The goal isn’t just faster automation — it’s resilient, continuously improving QA pipelines that adapt alongside the systems they’re testing. This shift means: ✅ Less manual maintenance. ✅ Higher reliability in CI/CD pipelines. ✅ More focus on quality insights, less on firefighting. We’re not just generating tests anymore. We’re building a living QA system that learns, adapts, and self-heals. 👉 What’s your view — will self-healing become the new baseline for enterprise test automation?

140

25 Comments

Here’s a truth I learned the hard way: QA isn’t the enemy of Devs. It’s their secret advantage. When we found 12 critical bugs before launch, the Dev team didn’t groan— they thanked us. Why? Because we caught what could have derailed their work. I’ve seen the memes. QA as villains... But reality looks different. • We protect product reputation • We accelerate go-live dates • We help Devs succeed The toughest part? Navigating pushback, not finding issues. But when trust builds, everyone wins. How do YOU transform friction into collaboration on your team? #QualityAssurance #Collaboration #SoftwareDevelopment

6

The QA Power Trifecta: From Idea to Flawless Release! 🎯 Building quality software isn't magic; it's a discipline that connects robust requirements to solid testing strategies within a defined lifecycle. Get any of these steps wrong, and your project stalls! I've synthesized four key visuals that form the blueprint for high-impact QA: 1. The SDLC Foundation 🏗️ The Software Development Life Cycle (SDLC) proves testing is not an afterthought. It's a stage (04) that directly follows Design and Coding, and feeds into Deployment, Monitoring, and Evaluation. The quality journey is continuous! 2. Defining Quality with Acceptance Criteria ✅ Quality starts with crystal-clear requirements. Use the Input-Process-Output (IPO) model to write User Story Acceptance Criteria. This manages expectations, reduces ambiguity for developers, and gives testers a definitive "Definition of Done." 3. Strategic Testing Mindset 🧠 Elevate your testing by following the 7 Strategies of a Better Software Tester. This means: Clearly Defining Objectives. Identifying Use Cases for all Users. Making Manual Exploratory Testing a habit. 4. Key Testing Methods 🧪 Your strategy needs execution. Understand the purpose of core types: Unit Testing (individual components), Integration Testing (modules working together), Security Testing (vulnerabilities), and Acceptance Testing (user sign-off). Mastering the connection between what you build (SDLC/Requirements) and how well you verify it (Testing Strategy) is the ultimate competitive advantage. What part of the process, Requirements, Testing, or SDLC is your team currently optimizing the most? Let me know! 👇 #SoftwareTesting #QATesting #SDLC #AcceptanceCriteria #Agile #TechStrategy #SoftwareDevelopment #Requirements

63

“It’s just a console warning.” “It doesn’t break anything.” “Let’s fix it later.” That was the initial response to a seemingly minor issue QA flagged — a 1-line fix. But that one line was the reason screen readers were skipping an entire form. It took: • 3 comments • 2 Slack threads • A 10-minute live demo to the stakeholder …before it clicked. That 1-line change made the platform accessible to thousands of users relying on assistive technologies. 💡 Here’s the hard truth: Testers often have to advocate fiercely for issues that don’t cause crashes — but cause exclusion. Being in QA isn’t just about finding bugs. It’s about speaking up for the users who can’t raise their voices through error logs. Let’s remember: ✅ A tiny fix can have a massive impact. ✅ Accessibility is not a nice-to-have — it’s a must. #AccessibilityMatters #QualityAdvocacy #InclusiveDesign #QARole #UserCenteredTesting #TestingWithEmpathy #SmallFixBigImpact #VoiceForUsers

20

I usually post about how important QA is after a major outage like there was this week. But that’s becoming so frequent I don’t want to bother constructing an entire post. So, here’s the main points: Stop laying people off, especially when profits are healthy. The workload is too much for many teams, morale plummets and things slip through. Stop looking at QA as an expense. It is a necessary part of development. One builds, one tests, result is a more reliable product. You can be a scion in tech just to become a meme laughingstock overnight. Also, entrepreneurs: it’s a great time to make better choices than those that are established and not protecting their teams and products. Reliabity matters. Guess I made a post after all. Sure it won’t be the last. #softwareengineering #qualitymatters #outages #layoffs #techleadership #entrepeneurs #startups #qa #engineeringleadership #productreliability

30

6 Comments

After 10+ years in QA, here’s the truth no one tells you: Not a single company cared about my degree or certifications. They cared about how I think, how I test, and how I communicate. They gave me practical tests, sometimes live, so they can evaluate how I think. And when I was the one hiring for teams I managed, I did the same. 📚 I have various trainings: 1. UX/UI 2. Security 3. ISTQB (Professional Tester) 4. Jira Admin 5. Scrum Master 6. Project Management 👩‍🎓 Plus a Bachelor’s degree in Computer Science. Literally, none asked in probably hundreds of interviews and chats with clients over the years. Are certifications worth it??? ✅ To break into QA — yes, 100% ❌ But collecting them just to impress recruiters? Not worth it. Focus on skills, not stickers. Focus on how to get the mindset. Focus on how to improve your skills. Focus on how you interact with the team. > Save this post if you’re starting out. > Share it with someone stuck in “certification collecting” mode. #RealTalkFromQA #SoftwareTesting #QAEngineer #QACommunity #QATips #QualityAssurance #TestAutomation #CareerAdvice #WomenInTech #TechTips #ManualTesting #QACertification #TestingMindset #LearnQA #TestLikeAPro #ScrumMaster #ISTQB #UXUI #JiraAdmin #WomenInQA

195

13 Comments

Construe, our configurable Med RAG API for code extraction, is definitely our most popular API on phenoml! We support 7 terminology systems on the market (ICD-10, LOINC, CPT, HPO, etc.) as well as custom coding systems (think adverse events, custom codes for payers). You can customize the exact combination of chunking, extraction, and validation to meet your specific use case. Examples of what we've seen developers build using Construe range from clinical decision support, tagging patient messages for personalization, enriching unstructured text for research, financial planning for clinical trials and billing/admin. Customers have asked us to add the ability to cite the exact text and we’re THRILLED to share that Construe now supports citation (e.g. citing the exact parts of the text that relate to the extracted code). Here’s a quick demo from our developer console of how it can be used for text highlighting! Coming soon for construe: built-in evals, tuning, and observability via API! Construe is available via API as well as with our Typescript (we use this in our console!), Python, and Java SDKs. Check it out! Developer console: https://console.pheno.ml developer docs: https://developer.pheno.ml

77

4 Comments

QA friends! Flaky tests are a universal axiom, intrinsic to any testing tool or framework. If someone tells you otherwise… well, you might want to reconsider who you listen to. 😄 Sure, some QA tools may be more prone to flaky tests than others, but flakiness is rarely caused by the tool itself. More often than not, it comes from environmental conditions, poor application design, or simply bad test design, usually when basic testing principles are not fully understood or the testing tool is being misused. 🌐 With the first one (the environment), there’s not much we can directly control. That said, good tests can help identify which environmental conditions are triggering failures, both in the application and in the tests themselves. ✏️ Regarding application design, tests can (or should) help uncover design flaws and edge cases that should be fixed. But priorities are usually decided elsewhere (Product Managers, Engineer Mangers, Product Owner, Leadership, you name it), and as QAs, we can’t fully guarantee the user experience, no matter how good our tests are. 🧠 The third one, though (test design) is the only area we truly control as QAs. You can learn, improve, and sharpen your testing principles. It takes time and effort, yes, but in the end, it is up to you. And if what you want is to really understand how your testing tool works, design better tests, and reduce (or even eliminate) flakiness caused by misuse… I can genuinely help with that, especially when we’re talking about Cypress.io. That is exactly why I built my brand-new open-source plugin: CYPRESS-FLAKY-TEST-AUDIT. It lets you see (visually and intuitively) what went wrong in your test: where it failed, the exact values passed to each Cypress command, and even side-by-side comparisons of different retries from the same test run to help you spot the root cause. GitHub: https://lnkd.in/g-WvJD_s npm: https://lnkd.in/gjWnUsue And if you want to truly understand Cypress’s async behavior and internal command queue, I have also written two public articles on the topic. They explain everything in a very intuitive way, without skipping the important details, so you can finally “get it” once and for all: - The Async Nature of Cypress: Don't Mess with the Timelines in Your Cypress Tests 'Dual-Verse': https://lnkd.in/gxA_WxXP - CYPRESS-FLAKY-TEST-AUDIT: thriving in the Cypress 'Dual-Verse' for once!: https://lnkd.in/gV7Y6ywY I’m not making empty promises, just check them out. And if after reading and using them you are not convinced, no worries… I’ll give you your money back. Oh wait... everything I share is free and open source 😅 So no money back… but I will apologize for wasting your time. Cheers! 🍻

69

7 Comments

After enough time in QA, you stop treating things as “working” or “not working”. You start noticing how they work. You double-check if the door is locked — not because you forgot, but because you don’t trust a single pass. An elevator takes a second too long to respond — you notice the lag. At a store, two identical products behave differently — inconsistent behavior. Someone explains something confidently, but with vague wording — low confidence in the output. QA is not a work, it’s quietly rewires how you observe the world. Once you’re trained to spot edge cases, delays, and ambiguity, you don’t switch it off after 6 PM.

184

26 Comments

Most people think the magic in autonomous testing is the recording. You click around, the tool records your actions, boom - test created. But here’s the truth: Recording is the easy part. The hard part is knowing what to test in the first place. Ask any QA leader - half the battle is just identifying the scenarios worth testing. It’s slow, it’s manual, and it demands deep product intuition. That’s why the real “Aha!” moment with SaolaAI hits when users realize: “Wait… I don’t need to think about that anymore.” Saola automatically surfaces the most critical flows - the ones your users actually perform. You don’t guess. You don’t assume. You see. Then comes the second “Aha!”: All of it runs without breaking - even when your UI changes. Because our second AI model doesn’t just replay steps - It decides, in real time, how and when to act. Like a human would.

129

6 Comments