As of 13 August 2026, AI can create a coding assessment for a job candidate.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the job description and write down the seniority, programming language, framework, essential skills and actual work the successful candidate will perform.
   2. Choose the assessment format, time limit, candidate environment, permitted resources and accessibility arrangements, then add these details to the prompt.
   3. Paste the prompt into a chatbot and save the candidate brief, evaluator guide, tests, reference solution and scoring rubric it produces.
   4. Copy the task into a clean repository or coding environment and run every example, edge-case test and reference solution yourself using the stated language and version.
   5. Ask a technically competent colleague who understands the role to solve the task without seeing the reference solution, then compare their approach and completion time with the proposed rubric.
   6. Rewrite ambiguous instructions, remove requirements unrelated to the role, record reasonable alternative solutions and send the final candidate brief and practical submission instructions.

   Prompt

   Create a fair, job-relevant coding assessment for this role.
   
   Role: [job title]
   Seniority: [entry, mid-level, senior or other]
   Programming language and version: [language and version]
   Frameworks, libraries or tools candidates may use: [list]
   Core skills to assess: [list]
   Skills that must not be assessed: [list]
   Time limit: [time limit]
   Assessment format: [take-home, live coding, pair programming or other]
   Expected candidate environment: [editor, repository, test runner and setup]
   Accessibility or reasonable-adjustment requirements: [requirements]
   Company or product context that may be used: [brief context]
   
   Produce an assessment pack containing:
   1. A candidate-facing brief with the objective, starting information, deliverables, assumptions, time limit, permitted resources and submission instructions.
   2. A task that tests the stated skills without relying on obscure trivia, unpaid work unrelated to the role, protected characteristics or knowledge of private company information.
   3. A small, unambiguous example input and output.
   4. A private evaluator's guide with functional requirements, edge cases, test cases and a reference solution.
   5. A scoring rubric with separate criteria for correctness, code quality, reasoning, testing and communication, with clear evidence for each score.
   6. Common acceptable approaches and reasonable alternative solutions.
   7. Likely failure modes, ways the wording could disadvantage candidates, and accessibility improvements.
   8. Three follow-up questions that test understanding without requiring candidates to guess your preferred implementation.
   
   Keep the task proportionate to the time limit. Do not invent company facts, libraries, APIs or requirements. Mark every assumption as an assumption. Make the candidate brief and evaluator guide separate. Do not present the draft as ready to use: end with a short checklist of items I must test or confirm before sending it to candidates.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.