Claude Certified Developer — Foundations, discussed: the full conversation as an article

More free episodes from this course

Full transcript

1. Morgan0:00

   If an answer choice sounds sensible, polished, and vaguely like something a senior person would say in a steering committee, should you trust it? Absolutely not. I’m Morgan.

2. Casey0:10

   I’m Casey, here to make sure “sounds sensible” is not mistaken for evidence. A surprising amount of exam technique is noticing that a sentence has arrived wearing a tie.

3. Morgan0:21

   Today we’re tackling the CCDV-F — the Anthropic Claude Certified Developer – Foundations certification — in this practical assessment lesson: reading questions, spotting distractors, deciding whether to change an answer, and interpreting the score.

4. Casey0:36

   It’s about stopping preventable losses under pressure. Which, annoyingly, is also most board meetings.

5. Casey0:42

   This podcast is brought to you by Memriq. While these courses focus on different aspects of the agentification of the enterprise, Memriq is focused on the agentification of your life: tools that remove the invisible administrative burden from everyday life, so you can focus on what you actually care about. Find out more at memriq dot A I.

6. Morgan1:03

   Here’s the hook: the question you lose may not cover unfamiliar material. It may offer several reasonable answers, one technically true but irrelevant, and another that is an elaborate solution to a problem nobody asked. A distractor is an incorrect answer option deliberately designed to compete with the correct one. The useful shift is to stop asking, “Could this be right?” and ask, “Is this the best answer to this exact request, at this exact layer, with this exact evidence?” That is the same discipline you use when evaluating suppliers, investments, or risk proposals. Today, we make it mechanical.

7. Casey1:39

   Though do not read a fixed shape into that. This paper mixes multiple-choice with multiple-response, so the number of options and the number of answers are whatever the item says they are.

8. Casey1:51

   TLDR: use five counters for five common wrong-answer shapes. Challenge universal claims with one counter-example. Reread the question’s final sentence. Identify the layer before judging the options. Prefer the simplest answer that fully meets the requirement. Restate the question before matching an answer. Change your choice only when you can name what you missed — never because your confidence wobbled. And CCDV-F uses a scaled score, meaning a result reported on a defined range rather than as a direct percentage: the pass mark is 720 on 100 to 1,000. It is not 72 percent, and it does not convert neatly into one.

9. Morgan2:30

   Let’s take Core Idea 1: the wrong-answer shapes to recognise. The point is not to distrust every option. It is to recognise recurring traps and know the counter that defeats each one.

10. Casey2:42

    The five are worth naming. The absolutist, which leans on always, never, all or only. True, but not the question. Right action, wrong layer. The over-engineered one. And the strong answer to a question nobody asked.

11. Morgan2:54

    One warning before any of that. The format is multiple-choice and multiple-response, so never assume four options and never assume a single answer. In the captured items twelve of seventeen offered four and five offered three — the item states what it wants, so read that line before you read the options.

12. Casey3:13

    Core Idea 2 is changing an answer you got on the first pass. The rule is severe because feelings are persuasive: change your answer only when you can name what you missed on the first read.

13. Morgan3:25

    So a change needs a receipt.

14. Casey3:27

    Exactly. Not a mood, a receipt. “I missed the word only.” “I overlooked the constraint.” “The question asks which action comes first.” Those are auditable reasons.

15. Morgan3:37

    Now Core Idea 3: the pass mark and how the number should change your preparation. CCDV-F — Anthropic Claude Certified Developer – Foundations — is a real published certification. Its pass mark is 720 on a scaled score running from 100 to 1,000.

16. Morgan3:54

    And the paper itself is fifty-three items in one hundred and twenty minutes, which is a shade over two minutes an item. That is comfortable for recall and workable for judgement, which is what this exam is. Developer stems run long, so budget your reading time, not just your deciding time.

17. Casey4:11

    So 720 is a threshold on the reporting scale, not a shopping list of questions to get right.

18. Morgan4:18

    The rest of the published facts are worth having in one place. It is one hundred and twenty-five dollars an attempt, there are no prerequisites, and the credential is awarded on exam performance alone. It lasts twelve months, and renewing on time is a free, non-proctored assessment; let it lapse and you sit the full exam again at full price.

19. Morgan4:39

    Precisely. That matters when you use practice results to make a decision about readiness. A raw correct-answer count may be useful feedback inside an exercise, but it is not the published pass mark. The pass mark is 720 scaled points.

20. Morgan4:54

    Think of an executive dashboard. A threshold on the dashboard is not interchangeable with the raw number of incidents behind it.

21. Casey5:02

    And “scaled” is not a secret bonus multiplier for optimism.

22. Morgan5:06

    It is not a curve either. The cut score is criterion-referenced — it comes from a formal standard-setting study, not from how the other candidates sitting that week happened to do.

23. Morgan5:17

    Treat 720 as the pass mark on 100 to 1,000, not as a percentage, and let practice performance guide study without inventing a conversion.

24. Casey5:27

    And your result report breaks the score out by domain. Those lines are commentary, not hurdles — scoring is on the total scaled score only, and there is no separate bar to clear in each domain.

25. Morgan5:39

    And if it goes badly, the retake waits run fourteen days, then thirty, then ninety, with the fee payable every time. The binding constraint is not the ladder though, it is four attempts per rolling twelve months. Sit when you are ready rather than early.

26. Casey5:54

    Core Idea 4 is why a wrong answer looks right. In a well-built certification item, wrong answers are written to compete with the right one. This is item construction: the deliberate design of the question and its choices. The distractor is not random debris. It is there because a knowledgeable person might select it for a specific reason.

27. Morgan6:15

    And the shapes repeat, so you can rehearse them. A short scenario with four candidate readings. A session trace where you are told what is not broken and asked for the surgical fix. A configuration with blanks to fill, deployment scenarios to match against transport and scope pairings, and a free-text item you answer before comparing with a model answer. The largest is a cumulative debug that plants four bugs, one each in the schema, streaming, context and memory layers, and asks you to name the layer as well as the bug.

28. Morgan6:47

    So the exam is testing judgement, not just memory.

29. Casey6:50

    Worth one caution, though. Everything we can see of how Anthropic writes these is courseware, not exam items — thirty-two of them, and not one carries a recorded correct answer. So anyone claiming to know the official answers is not reading them from the courseware.

30. Casey7:06

    Exactly. It tests whether you can connect knowledge to the stated objective and constraints. A technically true answer can still be wrong if it solves yesterday’s problem, operates at the wrong point in the system, or does more than the question requires.

31. Morgan7:22

    The developer version of that is a plausible-but-wrong API shape. The option names real fields in a real order and still repairs a different class of defect than the error message names — adding a required array to a schema when the real failure is a mismatched tool use id. Read the error first, decide which class it is, schema validation, message ordering, transport or authentication, then discard every option outside that class.

32. Morgan7:48

    “Not nonsense” is a very low bar.

33. Casey7:50

    A bar so low most distractors can limbo under it. The winning answer is the one that best matches the question, not the one that performs impressively in a different question.

34. Morgan8:01

    The nastiest developer one is an anti-pattern wearing best-practice clothing. Rotate the key and update the credentials file sounds disciplined and leaves the secret sitting in the file; concatenating every prior session transcript looks like conscientious memory and is the memory-layer bug. The tell is an option whose stated benefit is thoroughness or convenience while its cost is a control you have just removed.

35. Morgan8:27

    So that is the exam. Now the part I would not skip if I were you: how to use Alloy to pass it. Alloy is the course you are standing in right now, and over the next stretch of this show I am going to take you through the eight things Alloy does that a textbook and a question bank cannot, in the order you should actually do them. I want to be specific, because Alloy only pays you back if you use it the way it was built — and used properly it will hand you back weeks. Number one. Start with Alloy's diagnostic. Not after a warm-up chapter — first, before you study anything.

36. Casey9:01

    Before? Most people would want to revise a bit so they do not embarrass themselves.

37. Morgan9:07

    That instinct costs you a week. The diagnostic is not a quiz bolted on the side — it is the same engine, drawing from the same question bank, under the same rules as everything else you will do here. So what it tells you is where you stand against this exam, not against a warm-up.

38. Casey9:24

    And the mix of questions?

39. Morgan9:26

    Built to the exam's own shape. The areas it draws from, and how heavily it draws from each one, follow the way the exam itself is weighted — and inside an area it deliberately spreads across different tasks rather than asking you the same thing four times. So a weak spot in the results is not one bad question. It is proportionally expensive, and it tells you where your time is worth the most.

40. Morgan9:50

    And those weights are published, so you can read them before you sit anything. Applications and Integration is thirty-three point one percent, Model Selection and Optimization sixteen point eight, Agents and Workflows fourteen point seven, Prompt and Context Engineering eleven, Tools and MCPs ten point six, Security and Safety eight point one, Claude Code three point one, and Eval, Testing and Debugging two point six. Multiply each by fifty-three and you get a defensible expectation of where the items are — about eighteen in Applications and Integration, and about one in Eval.

41. Casey10:24

    So one domain is a third of the paper on its own, more than the two next-largest combined. And Claude Code and Eval together are five point seven percent — about three items — which makes them the two most enjoyable places to lose a third of your revision.

42. Morgan10:40

    This is also the one certification in the family whose guide publishes weights below the domain level — twenty-five weighted sub-objectives, each with its own share of the paper. Claude Application Design is the heaviest single one at eight point six percent, which on its own outweighs the entire Security and Safety domain. Claude Hooks is the lightest at one percent, roughly half an item — worth knowing what a hook is, not worth an evening.

43. Casey11:07

    How long are we talking?

44. Morgan11:09

    However long the course sets. And you cannot shorten it — the server will refuse. That is on purpose: a diagnostic that quietly gets smaller measures less and still reports the same confidence. Finish it in one sitting. A partial diagnostic is not a smaller diagnostic, it is a worse one.

45. Casey11:26

    What do I get at the end of it?

46. Morgan11:29

    A real starting number for how ready you are, and a breakdown by area on the summary screen. Read that breakdown while it is in front of you — that screen is the moment it is most useful to you. The time you spend here saves you a week of studying the wrong thing.

47. Morgan11:45

    And on the exam itself the same logic decides your plan. Because scoring is on the total, the rational move is to maximise total items correct — an hour that lifts your accuracy ten points in Applications and Integration is worth about one point eight items, while the same hour in Eval is worth about zero point one. Both are real improvements. One is worth about a dozen of the other.

48. Morgan12:08

    One sub-objective is worth calling out on its own. Software Engineering Foundations, at seven point four percent, is heavier than the largest sub-objective inside Tools and MCPs, which is Tool Implementation at four point four. Ordinary engineering discipline around a Claude application is worth more here than the Claude-specific mechanics most people spend their revision on.

49. Morgan12:31

    Second, and this is the one almost nothing else does. Every answer in Alloy asks you how confident you are, and Alloy will not let you submit without it.

50. Casey12:41

    I will be honest, my instinct is to click through that as fast as possible.

51. Morgan12:46

    Do not. That rating is the most valuable thing you give Alloy, and it is the one thing you cannot get back later. Alloy watches the gap between how sure you were and how right you turned out to be — and nothing else you do on this course tells Alloy that.

52. Casey13:01

    Why does the gap matter more than the score?

53. Morgan13:04

    Because being certain and wrong is what fails people. If you are unsure and wrong, you go back and check it. If you are certain and wrong, you never look at it again — and it walks straight into the exam with you. That gap is a trainable thing. It is also part of how ready you are judged to be, so if your knowledge looks fine and your readiness does not, the fix is honest ratings, not more questions.

54. Casey13:29

    And if I just rate everything at the top to be safe?

55. Morgan13:32

    Then you have thrown away the signal and bought nothing. Rate what you actually felt.

56. Morgan13:38

    Here is the payoff, and it is concrete: Alloy can hand you back exactly the questions you were confident about and wrong about. There is a practice mode for precisely that. Run it after your diagnostic. That set is the single highest-value pile of questions in your whole preparation, and it only exists if you rated honestly.

57. Morgan13:58

    Third. Alloy's flashcards, and the rule is: do them when Alloy says they are due, not when you feel like it. Do that and on exam day the facts are simply there, with no cram session anywhere in your week.

58. Casey14:10

    What is the difference?

59. Morgan14:12

    The timing is the whole trick, and getting it right is Alloy's job rather than yours. Each card comes back at the moment when recalling it makes it stick — which is just before you would have lost it. Do them early and you waste the review. Do them late and you have already forgotten it.

60. Casey14:29

    And it knows that how?

61. Morgan14:31

    From you. After each card it asks how well you recalled it, and you have four honest answers — again, hard, good, or easy. That answer is not a score and it is not stored as one. It is what decides when you see that card next.

62. Casey14:44

    So "again" is not a failure.

63. Morgan14:46

    "Again" is the most useful answer in the set. Rating yourself up to feel productive pushes the card away from you and costs you the review you needed. And do not cram a deck in one sitting — that throws the whole thing away. The queue is already in the right order when you open it. Work it from the top and stop choosing.

64. Morgan15:06

    Fourth, and this one is different from anything you have used before, so let me explain it properly — it is the reason Alloy's readiness number is worth trusting when other progress bars are not.

65. Casey15:18

    Go on.

66. Morgan15:19

    Most tools count what you clicked. Alloy weighs how much a thing actually proves. Reading a lesson is the weakest signal there is — you can read anything and retain none of it, and it says nothing about whether you could do it under pressure. Answering a question is stronger. Being right when you were confident is stronger still. Working a scenario or a lab is stronger again.

67. Casey15:42

    So "I read it" and "I could actually retrieve it" are not the same row in my record.

68. Morgan15:47

    They are not, and only the second one counts toward knowing something. Which is why nothing here is ever ticked off because you scrolled past it. You mark it, or you prove it. Marking a lesson complete is an explicit thing you do — and podcasts only count when you have genuinely listened to a real part of them, so skipping to the end credits you nothing.

69. Casey16:09

    That sounds like it makes my progress bar look worse.

70. Morgan16:12

    It makes it look TRUE. The readiness number you walk into the exam trusting is one that earned it. Inflating the strip only lies to you.

71. Morgan16:21

    One more thing you will see and should not misread. A concept can go quiet — it will tell you it needs a refresh. That is a fact about time passing, not a punishment and not a demotion. What fades is your ability to pull it back quickly, and the answer is one review, not a re-study.

72. Morgan16:38

    Fifth. When you sit down and you genuinely do not know what to do, open Alloy's Recommended next section and do what it tells you. That is the entire decision about what to study today, made and handed to you.

73. Casey16:51

    And I should just trust it?

74. Morgan16:53

    You should. What is sitting in Alloy's Recommended next section is the single action that moves your readiness the furthest today — it has already taken account of what you are about to forget, where you are weakest, and the heavy material you have not opened yet. That is the half hour you would have spent deciding, given back to you as a half hour of studying.

75. Casey17:15

    So it is not just picking the next chapter.

76. Morgan17:18

    Not remotely. A chapter list does not know you; Alloy does. It knows the question you got wrong last Tuesday while feeling sure of it, and it knows the topic you have not touched in weeks. Its order is not a menu — it is what would move your readiness most, right now. Do the one thing Alloy puts in front of you and you will never again wonder whether you are studying the right thing.

77. Morgan17:41

    Every element of Alloy is AI optimized to help you learn the most in the most effective and fastest way.

78. Casey17:48

    And when I do want to choose?

79. Morgan17:50

    Then choose by the question you are asking, because the practice modes are different questions. "Where am I thin" is one. "Where am I fooling myself" is another. "Can I hold it together for a whole paper" is a third. Pick the one that matches what you actually want to know.

80. Casey18:07

    The quick drill throws in things I already know, though.

81. Morgan18:10

    Deliberately. Mixing beats blocking, and a drill that only ever showed you your weakest material would collapse onto the same narrow set forever. The easy ones in there are the point, not padding.

82. Morgan18:22

    Sixth. Alloy gives you one readiness number, and unlike almost every progress bar you have ever seen, it is one you can actually act on.

83. Casey18:31

    What is behind it?

84. Morgan18:32

    More than raw accuracy — how well you actually know the material, how honest your confidence is, and how much of the course you have real evidence on. All three move it, so grinding more questions only moves one of them.

85. Casey18:46

    And a practice exam?

86. Morgan18:47

    Sit one when it tells you to. A real result adjusts your readiness in both directions, but it is bounded — one bad sitting cannot sink you and one good one cannot carry you. And the weight of that result fades as it ages, because how you did a month ago is not how ready you are today.

87. Morgan19:05

    On the real one, delivery is Pearson VUE, online-proctored or at a test centre. You accept the non-disclosure agreement before it begins — declining ends the session with no refund — and your government photo ID must be unexpired and carry the name on your registration exactly. Cancelling inside twenty-four hours forfeits the fee, as does not turning up, and any appeal must be filed within fourteen days.

88. Casey19:30

    So do not chase the number.

89. Morgan19:32

    Do not chase the number. Do the work it points at, and the number follows.

90. Morgan19:36

    One exception to ordering by weight. A domain where you are near zero rather than merely weak. Going from nothing to competent in Security and Safety is a bigger swing than its eight point one percent suggests, because security items are decidable once you know the failure modes and undecidable when you do not.

91. Morgan19:55

    Seventh, and this is the one I would actually put on the box.

92. Casey19:59

    Go on.

93. Morgan20:00

    These exams move. The tooling underneath them moves faster. And Alloy's curriculum is actively watched and kept current — not merely patched. New material is pulled into Alloy as the exam and the ecosystem change.

94. Casey20:13

    Which is not what happens to most prep material.

95. Morgan20:17

    Most prep material is written once and left. You buy it, and it slowly becomes a snapshot of what the exam used to be. That is worth saying plainly: what you study here is what the exam is now, not what it was when the course was written.

96. Casey20:31

    And that is people, not a machine.

97. Morgan20:34

    That is people. Someone watches these exams and pulls the changes in. It is a living curriculum, and it is the reason coming back is worth your time.

98. Morgan20:43

    Last one, quickly, because the rest of Alloy is worth knowing and each piece takes one breath to explain.

99. Morgan20:49

    And one honest absence. Nothing published says whether you can flag and revisit items, whether delivery is adaptive, whether partial credit or negative marking applies, or whether breaks and scratch paper are allowed. If your pacing plan depends on any of those, confirm it against the current exam specification when you register, rather than against advice.

100. Morgan21:11

     Labs — a lab step only appears when a real authored lab covers that concept. There is no filler here, so when you see one, it is worth doing. Open it when the concept is something you would have to DO rather than recall.

101. Morgan21:25

     Scenarios — scenario practice gives you the hard ones first, on purpose. Take one when the concept is a judgement call rather than a fact, because that is the form the exam tests judgement in.

102. Morgan21:37

     Your activity strip — it shows what you did and on which day, across the different kinds of work. It is a record of activity, not a coverage map, and a dash means that kind of work does not exist on this course rather than that you did none of it.

103. Morgan21:52

     And the Curriculum — that is the one map everything else is indexed against. Open it, find the concepts a lesson just covered, and you will see how the whole loop hangs together. Nothing you do here is disconnected from the exam's own shape.

104. Morgan22:06

     And read each domain the way the official material frames it. Every topic is there because it prevents a specific production failure, so ask which implementation decision prevents which failure. Read them that way and the scenario items stop looking like trivia.

105. Morgan22:22

     Twelve domain lessons follow this one, and they run in a teaching order rather than the blueprint's listing order — fundamentals before API mechanics, mechanics before what is built on them. Applications and Integration gets four of them because it is a third of the exam; every other domain gets one or two.

106. Casey22:41

     Do the diagnostic, rate honestly, keep the cards due, prove it rather than tick it, follow the Recommended next section, and come back because it keeps changing.

107. Morgan22:51

     And drill the reading order every time. The error or the evidence first, then the layer the defect lives in, then the options. Naming the layer before you look at the choices is what makes the trace items fast.

108. Morgan23:04

     Two more. Practise subtraction — one shape hands you four plausible pieces of configuration and tells you to leave out anything that does not belong, and that clause is the whole item. And review your reasoning rather than the letter you picked, because the right option for the wrong reason will not survive a rephrased trace.

109. Morgan23:24

     That is the whole thing.

110. Morgan23:26

     Next, we turn these recognition rules into repeatable performance with practice questions and exam simulations — so you can apply the checks quickly, under the same pressure that makes plausible wrong answers so expensive.

In writing