A PM asks what you would do in your first month as junior DevRel for a new payments API. What would you deliver for developers?

I would combine a small engineering artifact, useful content, and direct developer contact around the first payment journey. - Build and test a minimal TypeScript app that creates one sandbox payment. - Publish a quickstart from API key setup to a verified 200 response. - Host one office hour and classify the questions into docs, SDK, product, or support follow-ups.

Your API serves solo JavaScript developers and enterprise Java teams, but you can create only one onboarding asset this sprint. How would you segment the audience?

I would choose the segment with the clearest current onboarding gap instead of writing one generic asset for everyone. - Compare support questions, sign-ups, and first-call success for the JavaScript and Java cohorts. - Write a one-line need for each, such as local prototype in 10 minutes versus approved enterprise integration. - Pick one cohort with the PM and label the asset clearly so the other audience is not misled.

A new developer signs up for your image API but only 30 percent make a first successful call. How would you map the first-use journey?

I would map every required step from sign-up to a visible transformed image and look for the first large drop. - Record account creation, API key creation, SDK install, request sent, and successful response as separate steps. - Try the flow in a clean browser and empty project so cached credentials do not hide friction. - Pair event data with five session notes or support threads to explain why developers stop.

Three developers at a meetup say the authentication flow is confusing, while engineering says it works as designed. How would you advocate for the developers?

I would turn the complaints into reproducible evidence and bring engineering a specific friction point, not a vague demand. - Ask each developer where they stopped, which SDK they used, and what they expected next. - Reproduce the common path and capture the misleading screen, error, or missing step. - Share the pattern with impact, such as three of five testers failing before the first call, and propose a small fix.

After a beta launch you receive 25 comments across Discord and email. How would you classify the feedback for the PM?

I would classify each comment by the team that can act on it and preserve the developer's evidence. - Use buckets such as product request, confirmed defect, docs gap, SDK ergonomics, support case, and praise. - Add platform, SDK version, journey step, frequency, and severity instead of counting every comment equally. - Send the PM a short summary with top patterns plus links to the original quotes and reproductions.

A Python developer reports that your SDK fails during installation on Windows 11. What would a reproducible DX bug report contain?

I would give engineering the smallest setup that fails consistently on a clean Windows environment. - Record Windows, Python, pip, and SDK versions plus the exact install command. - Include expected behavior, actual terminal output, full error text, and whether a virtual environment was used. - Attach a tiny requirements file or repository and note the last working SDK version if known.

You are opening a public Discord for an open-source SDK with 200 early members. What should the code of conduct and moderation setup include?

I would publish short behavior rules and a private, predictable way to report violations before inviting more members. - Give examples of welcome behavior and banned harassment, discrimination, threats, and unwanted promotion. - Name the reporting channel, who can read reports, and possible actions from a warning to removal. - Train at least two moderators on the same incident log and escalation path so enforcement is consistent.

A demo endpoint fails during a live workshop for 40 developers, and the incident is not yet on the status page. How would you protect community trust?

I would acknowledge the failure immediately, avoid guessing, and give participants a useful fallback. - State what is known, save the request ID and timestamp, and alert the incident owner. - Switch to a recorded response or local mock so attendees can finish the exercise. - Send a follow-up with the confirmed cause, corrected sample, and status link when engineering publishes it.

You have ten minutes at a local Node.js meetup to introduce a new queue API. How would you choose the talk outcome?

I would target one runnable result: enqueue a job, process it, and show the completed status. - Start with the audience problem of moving slow work outside an HTTP request. - Use one small Node.js repository and avoid unrelated dashboard or architecture features. - End with a short URL, a working command, and one feedback question about the hardest step.

You can fix either a typo on a popular reference page or a broken API-key step blocking eight beta developers. Which do you choose and how do you explain it?

I would fix the blocked API-key step first because it prevents the core first-call journey. - Confirm the break in a clean account and estimate how many active beta users hit it. - Open a tracked issue with the reproduction and offer a temporary workaround. - Keep the typo in the backlog with page traffic and schedule it after the activation blocker.

Your quickstart should get a JavaScript developer from sign-up to the first successful weather API call in under ten minutes. What do you include?

I would include only the prerequisites and commands needed to produce one verifiable response. - State the Node.js version, sandbox account requirement, and where to create an API key. - Provide copyable install and run commands with a secure environment-variable placeholder. - Show the expected city and temperature fields, then test the steps in a clean directory.

Newsletter newcomers complete the storage API quickstart at 22%, while developers arriving from Discord complete it at 44%. You have one week to test whether channel-specific onboarding helps. What would you do?

I would run one bounded test for the lower-performing newsletter cohort instead of creating separate journeys for every channel. - Keep the current quickstart as the control and send half of newsletter newcomers to a shorter JavaScript path that begins with account creation. - Measure first-call success and median time to first call by assigned variant, using source and variant IDs without collecting code or secrets. - Invite five non-completers to explain where they stopped, then keep the variant only if activation improves without increasing support requests.

You need a demo app for frontend developers evaluating a real-time chat API. What would you build and deliberately leave out?

I would build one small browser chat that proves connect, send, receive, and reconnect behavior. - Use a familiar TypeScript framework, a clear README, and one command to start locally. - Keep server-issued temporary tokens so the permanent API secret never reaches the browser. - Leave out custom design systems, user profiles, and production scaling because they obscure the API journey.

A code sample for creating an invoice works once but cannot be reused by readers. What would you change before publishing it?

I would turn it into a small, repeatable example with explicit input, output, and cleanup behavior. - Read the API key from an environment variable and use clearly fake customer data. - Validate the response status and print the invoice ID instead of logging an unstructured object. - Add dependency versions and rerun from a fresh clone twice to catch hidden local state.

Nine partner developers ask conflicting questions about required fields for `POST /v1/messages` in the generated OpenAPI reference. The PM wants product evidence, not another guide. How would you aggregate the signal?

I would turn the questions into a small evidence set that Product and API Engineering can act on. - Tag each report by confusing field, source channel, developer stage, and outcome, then count how many developers are blocked before their first call. - Reproduce the two largest clusters in the sandbox and retain only sanitized schema excerpts, error codes, messages, and request IDs. - Send the PM and API owner one issue with impact, examples, and a proposed schema or error-message change, then report their decision back to the partners.

Most new sign-ups use TypeScript, but onboarding offers only curl. You can add one runnable sample this sprint. How would you treat the sample as an activation experiment?

I would test whether a minimal TypeScript sample improves first-call activation rather than judging it by views. - Build one pinned, runnable example that creates a sandbox resource and leaves authentication, retries, and production architecture out of scope. - Show it to half of eligible TypeScript sign-ups while the control keeps the curl path, then measure sample starts, successful calls, and median completion time. - Ask five developers who still fail where they stopped and keep, revise, or remove the sample based on activation and support evidence.

A first-time contributor submits a sample-app PR with a live test key and TLS verification disabled. A maintainer closes it as 'unsafe,' and the contributor says they will not return. How would you repair trust?

I would secure the repository first, then make the recovery respectful and specific so this is not their last contribution. - Ask the owner to rotate the exposed key, remove the unsafe branch from use, and avoid repeating credential details in public discussion. - Thank the contributor for the intended fix, explain the two concrete risks without blaming them, and point to the sandbox and TLS-safe pattern. - Offer a smaller follow-up issue with a maintainer pairing session, then credit the contributor publicly if the safe replacement is merged.

How would you prove that a five-step CLI quickstart is runnable before a launch to 500 newsletter subscribers?

I would verify it from a clean environment and make success machine-checkable. - Clone the published branch into a fresh container with only the documented runtime version. - Run every command exactly as shown, including account and environment-variable setup. - Add a final command or test that exits successfully only when the expected resource exists.

Five tutorial testers all receive `401 Unauthorized` at step three, but the sample token works for you. What do you investigate first?

I would suspect an undocumented credential or environment difference before rewriting the API explanation. - Compare token type, sandbox versus production host, required scope, and expiry with the testers. - Recreate a new account and token instead of relying on my older privileged credential. - Update the failing step with a credential check and the exact expected error for a wrong environment.

You must record a six-minute tutorial video for Python developers making their first transcription API call. How would you structure it?

I would show one complete result with minimal edits and a matching repository viewers can run. - Open with the finished transcript, then state the Python version, sandbox account, and source audio. - Zoom the terminal, explain each command briefly, and never display a real key or personal file path. - Add chapters, captions, repository link, and a pinned correction if the SDK changes after recording.

Developer Relations Engineer interview questions

100 real questions with model answers and explanations for Junior candidates.

See a Developer Relations Engineer resume example →

Practice with flashcards

Spaced repetition · Hunter Pass

Questions

developer-relations

I would combine a small engineering artifact, useful content, and direct developer contact around the first payment journey.

Build and test a minimal TypeScript app that creates one sandbox payment.
Publish a quickstart from API key setup to a verified 200 response.
Host one office hour and classify the questions into docs, SDK, product, or support follow-ups.

Why interviewers ask this: The interviewer checks whether you understand DevRel as engineering, content, and community work tied to one developer outcome.

onboardingjavascriptagile

I would choose the segment with the clearest current onboarding gap instead of writing one generic asset for everyone.

Compare support questions, sign-ups, and first-call success for the JavaScript and Java cohorts.
Write a one-line need for each, such as local prototype in 10 minutes versus approved enterprise integration.
Pick one cohort with the PM and label the asset clearly so the other audience is not misled.

Why interviewers ask this: A strong answer uses evidence and developer needs to define an audience rather than segmenting only by language.

first-use-journeyfirst-successful-call

I would map every required step from sign-up to a visible transformed image and look for the first large drop.

Record account creation, API key creation, SDK install, request sent, and successful response as separate steps.
Try the flow in a clean browser and empty project so cached credentials do not hide friction.
Pair event data with five session notes or support threads to explain why developers stop.

Why interviewers ask this: The interviewer evaluates whether you can connect a developer journey to both observable events and firsthand evidence.

meetupauthdesign

I would turn the complaints into reproducible evidence and bring engineering a specific friction point, not a vague demand.

Ask each developer where they stopped, which SDK they used, and what they expected next.
Reproduce the common path and capture the misleading screen, error, or missing step.
Share the pattern with impact, such as three of five testers failing before the first call, and propose a small fix.

Why interviewers ask this: A strong junior advocate translates community experience into evidence that product and engineering can act on.

discord-communityfeedback

I would classify each comment by the team that can act on it and preserve the developer's evidence.

Use buckets such as product request, confirmed defect, docs gap, SDK ergonomics, support case, and praise.
Add platform, SDK version, journey step, frequency, and severity instead of counting every comment equally.
Send the PM a short summary with top patterns plus links to the original quotes and reproductions.

Why interviewers ask this: The interviewer checks whether you can convert noisy feedback into traceable product signal without flattening its severity.

reproducibilitypythonbug-reporting

I would give engineering the smallest setup that fails consistently on a clean Windows environment.

Record Windows, Python, pip, and SDK versions plus the exact install command.
Include expected behavior, actual terminal output, full error text, and whether a virtual environment was used.
Attach a tiny requirements file or repository and note the last working SDK version if known.

Why interviewers ask this: A strong answer makes the developer experience failure repeatable rather than forwarding a screenshot with no environment details.

community-moderationcode-of-conductsdk

I would publish short behavior rules and a private, predictable way to report violations before inviting more members.

Give examples of welcome behavior and banned harassment, discrimination, threats, and unwanted promotion.
Name the reporting channel, who can read reports, and possible actions from a warning to removal.
Train at least two moderators on the same incident log and escalation path so enforcement is consistent.

Why interviewers ask this: The interviewer evaluates whether you treat community safety as an operating process rather than a link in the footer.

workshopendpointsincidents

I would acknowledge the failure immediately, avoid guessing, and give participants a useful fallback.

State what is known, save the request ID and timestamp, and alert the incident owner.
Switch to a recorded response or local mock so attendees can finish the exercise.
Send a follow-up with the confirmed cause, corrected sample, and status link when engineering publishes it.

Why interviewers ask this: A strong response preserves trust through accuracy, ownership, and follow-up instead of pretending the demo succeeded.

data-structuresmeetup

I would target one runnable result: enqueue a job, process it, and show the completed status.

Start with the audience problem of moving slow work outside an HTTP request.
Use one small Node.js repository and avoid unrelated dashboard or architecture features.
End with a short URL, a working command, and one feedback question about the hardest step.

Why interviewers ask this: The interviewer checks whether you can scope advocacy content to an audience, artifact, and achievable developer outcome.

api

I would fix the blocked API-key step first because it prevents the core first-call journey.

Confirm the break in a clean account and estimate how many active beta users hit it.
Open a tracked issue with the reproduction and offer a temporary workaround.
Keep the typo in the backlog with page traffic and schedule it after the activation blocker.

Why interviewers ask this: A strong answer prioritizes developer impact while keeping lower-severity work visible and traceable.

javascriptquickstart

I would include only the prerequisites and commands needed to produce one verifiable response.

State the Node.js version, sandbox account requirement, and where to create an API key.
Provide copyable install and run commands with a secure environment-variable placeholder.
Show the expected city and temperature fields, then test the steps in a clean directory.

Why interviewers ask this: The interviewer evaluates whether you can design a quickstart around time to first value rather than feature coverage.

developer-onboardingquickstartapi

I would run one bounded test for the lower-performing newsletter cohort instead of creating separate journeys for every channel.

Keep the current quickstart as the control and send half of newsletter newcomers to a shorter JavaScript path that begins with account creation.
Measure first-call success and median time to first call by assigned variant, using source and variant IDs without collecting code or secrets.
Invite five non-completers to explain where they stopped, then keep the variant only if activation improves without increasing support requests.

Why interviewers ask this: The interviewer checks whether you can adapt onboarding to one audience, test it, and use developer evidence to decide what stays.

api

I would build one small browser chat that proves connect, send, receive, and reconnect behavior.

Use a familiar TypeScript framework, a clear README, and one command to start locally.
Keep server-issued temporary tokens so the permanent API secret never reaches the browser.
Leave out custom design systems, user profiles, and production scaling because they obscure the API journey.

Why interviewers ask this: The interviewer checks whether you can build a focused, secure engineering artifact instead of a polished but confusing showcase.

code-sample

I would turn it into a small, repeatable example with explicit input, output, and cleanup behavior.

Read the API key from an environment variable and use clearly fake customer data.
Validate the response status and print the invoice ID instead of logging an unstructured object.
Add dependency versions and rerun from a fresh clone twice to catch hidden local state.

Why interviewers ask this: A strong answer treats a sample as runnable code that developers can verify and adapt safely.

openapiopenapi-governanceaggregation

I would turn the questions into a small evidence set that Product and API Engineering can act on.

Tag each report by confusing field, source channel, developer stage, and outcome, then count how many developers are blocked before their first call.
Reproduce the two largest clusters in the sandbox and retain only sanitized schema excerpts, error codes, messages, and request IDs.
Send the PM and API owner one issue with impact, examples, and a proposed schema or error-message change, then report their decision back to the partners.

Why interviewers ask this: The interviewer evaluates whether you can convert repeated OpenAPI confusion into verified product feedback and close the community loop.

activationdeveloper-activationtypescript

I would test whether a minimal TypeScript sample improves first-call activation rather than judging it by views.

Build one pinned, runnable example that creates a sandbox resource and leaves authentication, retries, and production architecture out of scope.
Show it to half of eligible TypeScript sign-ups while the control keeps the curl path, then measure sample starts, successful calls, and median completion time.
Ask five developers who still fail where they stopped and keep, revise, or remove the sample based on activation and support evidence.

Why interviewers ask this: A strong answer treats a code sample as a testable adoption intervention with a narrow outcome and a clear decision.

contributorsmaintainerstls

I would secure the repository first, then make the recovery respectful and specific so this is not their last contribution.

Ask the owner to rotate the exposed key, remove the unsafe branch from use, and avoid repeating credential details in public discussion.
Thank the contributor for the intended fix, explain the two concrete risks without blaming them, and point to the sandbox and TLS-safe pattern.
Offer a smaller follow-up issue with a maintainer pairing session, then credit the contributor publicly if the safe replacement is merged.

Why interviewers ask this: The interviewer checks whether you can protect developers and rebuild contributor trust after an unsafe pull request.

quickstartnewsletter

I would verify it from a clean environment and make success machine-checkable.

Clone the published branch into a fresh container with only the documented runtime version.
Run every command exactly as shown, including account and environment-variable setup.
Add a final command or test that exits successfully only when the expected resource exists.

Why interviewers ask this: A strong answer replaces author confidence with repeatable verification under the reader's starting conditions.

tutorialtokens

I would suspect an undocumented credential or environment difference before rewriting the API explanation.

Compare token type, sandbox versus production host, required scope, and expiry with the testers.
Recreate a new account and token instead of relying on my older privileged credential.
Update the failing step with a credential check and the exact expected error for a wrong environment.

Why interviewers ask this: The interviewer evaluates whether you debug the reader's real setup and remove hidden author privileges.

pythontutorial

I would show one complete result with minimal edits and a matching repository viewers can run.

Open with the finished transcript, then state the Python version, sandbox account, and source audio.
Zoom the terminal, explain each command briefly, and never display a real key or personal file path.
Add chapters, captions, repository link, and a pinned correction if the SDK changes after recording.

Why interviewers ask this: A strong answer combines a focused developer journey with accessible, secure, and maintainable video basics.

Locked questions

21
A Discord member posts the same SDK error in four channels and pings maintainers repeatedly. How would you moderate it?
discord-communitymaintainerssdk
22
A GitHub Discussion asks whether an unsupported mobile platform will be added next quarter. How would you respond?
github-community
23
You find a Stack Overflow question about your API with an accepted but insecure answer that hardcodes a secret. What do you do?
secretsmemory
24
Only three developers attend a scheduled office hour, and all are blocked on webhook setup. How would you run and follow up on the session?
sessionswebhooks
25
During a livestream, one participant directs a discriminatory insult at another. What is your immediate escalation path?
escalationlivestream
26
Your community dashboard shows 18 unanswered technical questions older than 48 hours. How would you reduce the queue?
developer-communitydata-structures
27
A community member submits their third useful fix to your sample repository. How would you recognize the contribution?
developer-communitysample-repository
28
A developer sends you database credentials in a Discord direct message and asks you to debug their account. How do you maintain community boundaries?
discord-communitydeveloper-communitydatabase
29
The same pagination question appears weekly in GitHub Discussions. What would you do beyond answering it again?
github-communitypagination
30
Discord users request dark mode, while GitHub users report failed SDK retries. How would you summarize both signals for the product team?
sdkgithub-communitytheming
31
A developer's Node.js SDK call returns 401 although the equivalent curl request succeeds. How would you troubleshoot it?
http-toolssdk
32
A sample app starts returning 429 after a developer loops over 200 records. What guidance would you give?
sample-app
33
A webhook appears as delivered in the provider dashboard, but the developer's local app never processes it. Walk through your debugging steps.
webhooksconcurrency
34
You can reproduce a crash in SDK version 2.4.0 and need to open a public GitHub issue. What do you include?
sdkgithub-community
35
A 300-line Next.js demo fails when uploading files through your API. How would you create a minimal reproduction for engineering?
nextjs
36
SDK 3.2 changes retry behavior and removes a deprecated method, and 14 community members are stuck upgrading. You can run one 45-minute migration clinic. How would you use it?
developer-communitymigrationsproblem-solving
37
Community installation reports span Python 3.9, 3.10, and 3.13 on Windows and Linux, while the SDK team supports Python 3.10 through 3.12. How would junior DevRel create a support-matrix feedback loop without redefining the policy?
feedbackfeedback-looppython
38
Five customers report duplicate orders after the SDK retries timed-out POST requests. How would you hand this product defect to engineering?
defectssdk
39
Six community libraries wrap `createUpload`, and a proposed SDK rename to `uploads.create` may break them next month. The PM asks junior DevRel for input to the RFC. What would you contribute?
developer-relationsdecision-makingsdk
40
Developers report API 500 errors in Discord during a release, but engineering has not confirmed an incident. What do you communicate?
communicationincidentsapi
41
A PM asks for a simple activation funnel for the sandbox API. Which steps would you measure first?
funnelactivationdeveloper-activation
42
You are instrumenting a quickstart in PostHog or Plausible. What events would you add without collecting code or secrets?
quickstartsecretsposthog
43
Median time to first API call rose from 8 to 19 minutes after a new authentication screen launched. How would you investigate?
authtime-to-first-call
44
A tutorial gets 2,000 views but only 120 verified completions. What would you examine before calling it unsuccessful?
content-completiontutorial
45
A ten-minute webhook video loses 55 percent of viewers in the first 45 seconds. What would you change and measure?
webhooks
46
Twelve developers attend office hours, but attendance alone does not show value. Which outcomes would you record?
office-hours
47
You can send either a monthly general newsletter or a targeted SDK upgrade email to 800 active Python developers. How would you choose?
newsletterpythonsdk
48
You have budget for either a 100-person local workshop or four online office hours. What basic evidence would guide the choice?
workshopoffice-hours
49
Seven of twenty tutorial testers ask for OAuth before they can finish a partner integration. How would you close the feedback loop with the PM?
feedbackfeedback-loopoauth
50
After simplifying an API-key quickstart, first-call conversion rises from 32 to 41 percent. What would you do next?
quickstart
51
PostHog shows quickstart completion fell from 41% to 24% after Tuesday's docs release, and 17 sessions stop at the first API call; what would you do?
quickstartdocsapi
52
Twelve developers report redirect_uri_mismatch while following a 10-minute OAuth quickstart; how would you repair the onboarding path?
developer-onboardingquickstartoauth
53
An export API demo polls forever when a job enters failed status, and 14 developers abandon the sample; how would you repair it?
api
54
A webhook demo returns 401 for every signed event because its example reads the parsed JSON body before verification; how would you fix and prove it?
webhooks
55
You notice a real sk_live key in a public demo repository that has been visible for 35 minutes; what do you do first?
56
After you deploy a fixed browser demo, six testers still load JavaScript that calls the retired endpoint; how would you investigate?
deploymentendpointsjavascript
57
A browser quickstart succeeds with curl but fails with a CORS error for 19 developers using localhost:5173; how would you respond?
quickstartcorshttp-tools
58
After a package release, the JavaScript demo fails with ERR_REQUIRE_ESM on Node.js 20 while weekly successful runs drop from 320 to 190; what would you do?
zero-to-onejavascript
59
During a browser-upload workshop, 16 of 20 attendees receive 415 while the equivalent curl multipart request works. Twenty-five minutes remain. How would you recover the session?
workshopsessionshttp-tools
60
A community pull request simplifies a webhook sample by replacing constant-time signature comparison with ==; how would you review it?
developer-communitywebhookscode-review
61
A Discord help channel receives 46 identical promotional links from three new accounts in 20 minutes; how would you moderate it?
discord-community
62
At office hours, a developer asks whether an undocumented endpoint can be used in production, and no API engineer is present; what would you say?
endpointsoffice-hours
63
A newcomer asks where to create a sandbox key, and an experienced Discord member replies only 'read the docs'; how would you handle it?
docsdiscord-community
64
Two experienced members start insulting each other in Discord over whether REST or GraphQL is better, and 30 people are watching; how would you intervene?
restgraphqldiscord-community
65
A Stack Overflow question claims your SDK silently deletes records, but the code sample calls deleteAll() inside a cleanup block; how would you answer?
sdkcode-samplememory
66
A user posts a possible authentication bypass with request headers and tokens in a public GitHub issue; what would you do?
authtokensgithub-community
67
Thirty Spanish-speaking developers ask for a separate Discord channel, but the moderation team cannot cover another space yet; what would you recommend?
discord-communitycommunity-moderation
68
Seven GitHub issues report the same Windows installation error under different titles; how would you handle them?
github-community
69
A member who already received a warning and temporary mute for harassment now sends hostile direct messages to newcomers; what would you do?
70
A respected Discord member recommends disabling TLS verification to fix local SDK errors, and newcomers begin copying the workaround; how would you respond?
discord-communitytlssdk
71
You are asked to run a 45-minute beginner workshop on the JavaScript SDK for 40 registrants; how would you define and measure success?
workshopjavascriptsdk
72
In a beginner workshop, advanced attendees dominate Q&A while eight beginners cannot finish the first authenticated call; how would you recover the session?
workshopsessions
73
You have a 12-minute live demo tomorrow, but the staging API failed twice during rehearsal; how would you prepare?
api
74
A launch partner asks you to demonstrate an untested beta endpoint as officially supported in tomorrow's webinar; what would you do?
endpoints
75
A live tutorial ends with 18 unanswered Q&A messages, including 4 about data residency that you cannot answer; what would you do next?
tutorial
76
Fifty-two of 120 registrants attend your webinar, 31 stay to the end, and 9 run the sample within 48 hours; how would you report it?
77
Ten minutes into a 30-person hands-on workshop, 11 attendees are still installing dependencies; how would you recover the session?
workshopdependenciessessions
78
Your recorded tutorial is scheduled for Friday, but automated captions mis-transcribe 23 API and SDK terms; what would you do?
sdktutorial
79
Only 6 of 70 workshop attendees answer a 12-question survey; how would you improve feedback collection next time?
workshopfeedback
80
A tutorial livestream drew 95 attendees from 180 registrations, but developers in Asia submitted 14 questions after it ended; what follow-up would you run?
tutoriallivestream
81
Nine developers ask for a skipSignatureVerification option because local webhook testing is difficult; how would you route that feedback?
feedbackwebhookstesting
82
Three developers say the Go SDK 'sometimes hangs' during upload; what reproduction would you write before routing it?
sdk
83
A TypeScript sample cannot compile because the SDK response type omits retryAfter even though the field exists in OpenAPI; where would you route it?
openapiopenapi-governancetypescript
84
curl and all three SDKs receive 502 for POST /imports with files above 20 MB; how would you route the feedback?
feedbackhttp-tools
85
Nine users request a dashboard button for rotating API keys, while rotation already works through the CLI; how would you route this feedback?
feedback
86
A public GitHub issue you reproduced is fixed in SDK 3.2.1 after 24 developers subscribed; how would you close the loop?
sdkgithub-community
87
After a dashboard redesign, five community reports say token creation is broken. Three mention the old menu label, while two reach the page but cannot see the button. How would you classify and route the feedback?
developer-communityfeedbacktokens
88
The API and docs are correct, but 14 workshop attendees send timestamps in local format instead of RFC 3339; what kind of gap is this?
decision-makingdocsworkshop
89
Support reports 22 'webhooks do not work' tickets, but the tickets may cover several causes; how would you investigate the cluster?
webhooks
90
A developer shares a production log containing customer email addresses while reporting an SDK error; how would you preserve useful feedback safely?
feedbacksdk
91
PostHog shows daily quickstart_started events steady at 400, but api_call_succeeded falls from 180 to 70 overnight; what would you check first?
quickstartposthog
92
The quickstart_completed PostHog event fires when users open the final page, even if they never make an API call; how would you improve the measurement?
quickstartposthog
93
A tutorial receives 2,000 visits from three partner links, but only 44 sample starts have a known source; how would you improve attribution?
program-attributiontutorial
94
You can send one of two newsletter subject lines to 4,000 subscribers to promote a Python tutorial; how would you run the experiment?
experimentspythontutorial
95
For the next beginner event, you want to test whether sending a starter repository 48 hours early improves completion; how would you set it up?
content-completion
96
A developer opens their first open-source issue asking to contribute, but the repository has 140 open issues and no obvious starting point; how would you help?
97
A first-time contributor submits a useful 180-line pull request, but CI fails formatting and one test; how would you respond?
contributorscode-review
98
Within 24 hours of SDK 5.1, GitHub receives 16 comments about a renamed upload method that appears only in a long changelog; what would you do?
sdkgithub-community
99
A sample repository gained 600 GitHub stars after a launch, but weekly successful API calls stayed near 90; how would you interpret that?
github-communitysample-repositoryapi
100
After SDK 2.8 ships, activation for new Ruby users falls from 36% to 27% while Python remains at 35%; how would you gather release feedback?
activationdeveloper-activationfeedback

Questions

1A PM asks what you would do in your first month as junior DevRel for a new payments API. What would you deliver for developers?

2Your API serves solo JavaScript developers and enterprise Java teams, but you can create only one onboarding asset this sprint. How would you segment the audience?

3A new developer signs up for your image API but only 30 percent make a first successful call. How would you map the first-use journey?

4Three developers at a meetup say the authentication flow is confusing, while engineering says it works as designed. How would you advocate for the developers?

5After a beta launch you receive 25 comments across Discord and email. How would you classify the feedback for the PM?

6A Python developer reports that your SDK fails during installation on Windows 11. What would a reproducible DX bug report contain?

7You are opening a public Discord for an open-source SDK with 200 early members. What should the code of conduct and moderation setup include?

8A demo endpoint fails during a live workshop for 40 developers, and the incident is not yet on the status page. How would you protect community trust?

9You have ten minutes at a local Node.js meetup to introduce a new queue API. How would you choose the talk outcome?

10You can fix either a typo on a popular reference page or a broken API-key step blocking eight beta developers. Which do you choose and how do you explain it?

11Your quickstart should get a JavaScript developer from sign-up to the first successful weather API call in under ten minutes. What do you include?

12Newsletter newcomers complete the storage API quickstart at 22%, while developers arriving from Discord complete it at 44%. You have one week to test whether channel-specific onboarding helps. What would you do?

13You need a demo app for frontend developers evaluating a real-time chat API. What would you build and deliberately leave out?

14A code sample for creating an invoice works once but cannot be reused by readers. What would you change before publishing it?

15Nine partner developers ask conflicting questions about required fields for `POST /v1/messages` in the generated OpenAPI reference. The PM wants product evidence, not another guide. How would you aggregate the signal?

16Most new sign-ups use TypeScript, but onboarding offers only curl. You can add one runnable sample this sprint. How would you treat the sample as an activation experiment?

17A first-time contributor submits a sample-app PR with a live test key and TLS verification disabled. A maintainer closes it as 'unsafe,' and the contributor says they will not return. How would you repair trust?

18How would you prove that a five-step CLI quickstart is runnable before a launch to 500 newsletter subscribers?

19Five tutorial testers all receive `401 Unauthorized` at step three, but the sample token works for you. What do you investigate first?

20You must record a six-minute tutorial video for Python developers making their first transcription API call. How would you structure it?

Locked questions