<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Exametics Exam Prep]]></title><description><![CDATA[Exametics Exam Prep]]></description><link>https://exametics.hashnode.dev</link><image><url>https://cdn.hashnode.com/uploads/logos/6af80d6a97fffc20d7ebc4f6/2d77e3c9-de1f-4854-8664-7cf6ff328b9a.png</url><title>Exametics Exam Prep</title><link>https://exametics.hashnode.dev</link></image><generator>RSS for Node</generator><lastBuildDate>Sun, 22 Nov 2026 16:33:32 GMT</lastBuildDate><atom:link href="https://exametics.hashnode.dev/rss.xml" rel="self" type="application/rss+xml"/><language><![CDATA[en]]></language><ttl>60</ttl><item><title><![CDATA[AI-900 is gone. The exam is AI-901 now, and the difference is generative AI.]]></title><description><![CDATA[If you bookmarked a study plan for AI-900, it is out of date. Microsoft retired that exam on 30 June 2026 and replaced it with AI-901. The certification…]]></description><link>https://exametics.hashnode.dev/ai-900-is-gone-the-exam-is-ai-901-now-and-the-difference-is-generative-ai</link><guid isPermaLink="true">https://exametics.hashnode.dev/ai-900-is-gone-the-exam-is-ai-901-now-and-the-difference-is-generative-ai</guid><category><![CDATA[Azure]]></category><category><![CDATA[AI]]></category><dc:creator><![CDATA[Exametics]]></dc:creator><pubDate>Wed, 18 Nov 2026 12:39:29 GMT</pubDate><content:encoded><![CDATA[<p>If you bookmarked a study plan for <strong>AI-900</strong>, it is out of date. Microsoft retired that exam on <strong>30 June 2026</strong> and replaced it with <strong>AI-901</strong>. The certification name has not changed — it is still Microsoft Certified: Azure AI Fundamentals — only the exam behind it.</p>
<p>This matters more than a code change usually does, because the replacement is not a cosmetic refresh.</p>
<h2>What AI-901 actually covers</h2>
<p>From Microsoft's own exam page, AI-901 asks you to describe:</p>
<ul>
<li>Artificial Intelligence workloads and considerations</li>
<li>Fundamental principles of machine learning on Azure</li>
<li>Features of computer vision workloads on Azure</li>
<li>Features of Natural Language Processing (NLP) workloads on Azure</li>
<li><strong>Features of generative AI workloads on Azure</strong></li>
</ul>
<p>That last line is the point. Generative AI is now a first-class domain of the fundamentals exam, not an aside. If your study material predates mid-2026, it was written for a syllabus where that domain did not carry its own weight.</p>
<h2>What this means in practice</h2>
<p><strong>Old practice questions are the trap, not old concepts.</strong> Responsible AI, supervised versus unsupervised learning, the difference between a classification and a regression workload — none of that has gone anywhere. What has changed is the share of the paper you should expect on generative AI: prompts, grounding, the Azure services that host models, and the considerations around deploying them.</p>
<p><strong>"AI-900 dumps" are now actively misleading.</strong> Any question bank still labelled AI-900 was built against a retired blueprint. It will under-represent exactly the domain you are least likely to have picked up on the job.</p>
<p><strong>Check your voucher.</strong> If you bought an exam voucher before the switch, confirm which exam it books. Microsoft's guidance is that attempts must be made before the retirement date to count — after it, you sit the replacement.</p>
<h2>Two details worth knowing</h2>
<p>AI-901 currently shows <strong>no retirement date</strong>, so it is the stable target for the foreseeable future. And it is offered in thirteen languages, including Arabic, Japanese, Korean, Portuguese (Brazil) and both Simplified and Traditional Chinese — useful if English is not your first language and you have been assuming you had no choice.</p>
<h2>A general rule this taught me</h2>
<p>Exam codes do not always move in the direction that looks obvious. A lower number is not automatically the older one, and a retired exam's page still loads fine — it just carries a warning box naming the replacement. Read the vendor's page before you trust any article, including this one.</p>
<hr />
<p>Our full breakdown of the AI-901 syllabus, including what each domain actually tests, is here: <a href="https://exametics.com/ai-901-exam-guide/">AI-901 exam guide</a></p>
]]></content:encoded></item><item><title><![CDATA[AWS SAA-C03: 720 is not 72%, and two other things about the scoring]]></title><description><![CDATA[Three things about AWS SAA-C03 scoring that study groups routinely get wrong. All three come straight from AWS's own exam guide. AWS reports SAA-C03 on a…]]></description><link>https://exametics.hashnode.dev/aws-saa-c03-720-is-not-72-and-two-other-things-about-the-scoring</link><guid isPermaLink="true">https://exametics.hashnode.dev/aws-saa-c03-720-is-not-72-and-two-other-things-about-the-scoring</guid><dc:creator><![CDATA[Exametics]]></dc:creator><pubDate>Sat, 14 Nov 2026 07:48:53 GMT</pubDate><content:encoded><![CDATA[<p>Three things about AWS SAA-C03 scoring that study groups routinely get wrong. All three come straight from AWS's own exam guide.</p>
<h2>720 is a scaled score</h2>
<p>AWS reports SAA-C03 on a 100-1,000 scale and passes you at 720. Scaled scoring equates results across exam forms of differing difficulty, so there is no fixed number of questions you must get right, and AWS publishes no conversion. <strong>720 is not 72%.</strong></p>
<h2>15 of the 65 questions do not count</h2>
<p>AWS scores 50 questions. The other 15 are unscored items being trialled for future exams, and they are not identified. Practical upshot: an unfamiliar question may simply be one of the 15, so do not let it eat your clock.</p>
<h2>Scoring is compensatory</h2>
<p>You pass on the overall score, not per domain. That matters because the weighting is uneven:</p>
<table>
<thead>
<tr>
<th>Domain</th>
<th>Weight</th>
</tr>
</thead>
<tbody><tr>
<td>Design Secure Architectures</td>
<td>30%</td>
</tr>
<tr>
<td>Design Resilient Architectures</td>
<td>26%</td>
</tr>
<tr>
<td>Design High-Performing Architectures</td>
<td>24%</td>
</tr>
<tr>
<td>Design Cost-Optimized Architectures</td>
<td>20%</td>
</tr>
</tbody></table>
<p>Strength in the 30% domain genuinely offsets weakness in the 20% one.</p>
<p>One more: unanswered questions score as incorrect and there is <strong>no penalty for guessing</strong>. Never leave a blank.</p>
<hr />
<p>The practical consequence of all three is the same: a round percentage on a practice test is the wrong target. Weighted, domain-level coverage is the right one.</p>
<p>I wrote the longer version, with the reasoning and the AWS source links, over on our blog: <strong><a href="https://exametics.wordpress.com/2026/11/14/aws-saa-c03-720-is-not-72-and-two-other-things-about-the-scoring/">AWS SAA-C03: 720 is not 72%, and two other things about the scoring</a></strong></p>
]]></content:encoded></item></channel></rss>