How do you hire AI-native engineers?

You hire AI-native engineers by assessing them on real engineering tasks with AI turned on and reviewing how each candidate directs and checks the AI. In Codility, candidates first complete an AI-native task in Screen. Finalists then build with Claude Code in a live session in Interview.

Which coding assessment platform has AI-native tasks for agentic engineering roles?

Codility runs AI-native tasks in Screen, Interview and Skills Intelligence. Build AI tasks ask engineers to build an AI system, such as a pipeline that answers questions from a document archive. Work with AI tasks suit any engineering role, including roles where engineers work with a coding agent such as Claude Code.

How do you test whether an engineer can build with AI?

Give the engineer a realistic task with AI available. Review the output alongside the AI activity, then ask the engineer to explain the changes they made to the AI's suggestions. In Codility, your team can review the candidate report and discuss those decisions in a live interview.

Does Codility use AI to score engineers?

No. Scoring follows fixed rules against test cases, so the same output always gets the same score. Your team reviews the results and decides who advances.

Hire AI-native engineers who catch AI errors

Codility’s skills assessment platform tests how candidates and current engineers build with AI in realistic VS Code tasks. Review their output and AI activity, then discuss their decisions in a live interview.

Request a free trial

Order export API Work with AI · python

AI Copilot on

42:18 Submit task

export.py test_export.py

                  123456789101112131415
                

                  import math
                  from fastapi import APIRouter, Query
                  from .db import count_orders, fetch_orders
                   
                  router = APIRouter()
                   
                  @router.get(“/orders/export”)
                  def export_orders(page: int = Query(1, ge=1),
                                    size: int = Query(100, ge=1, le=500)):
                      total = count_orders()
                      offset = (page – 1) * size
                      rows = fetch_orders(offset=offset, limit=size)
                      pages = math.ceil(total / size)
                      return {“page”: page, “pages”: pages,
                              “items”: rows}
                   
                

TerminalProblemsOutput
candidate@orders-service$ pytest -q….                                                   [100%]4 passed in 0.38scandidate@orders-service$ 

AI Copilot

Activity is recorded

The last page of orders is missing from the export. Find the cause.

pages uses floor division, so a partial last page is dropped. Suggested fix:

pages = total // size + 1

Rejected by candidate · 04:31

That adds an empty page when total is a multiple of size. Use ceil and add a test for total=500, size=100.

Updated line 13 and added test_exact_multiple.

pages = math.ceil(total / size)

Accepted by candidate · 06:22

Ask the AI Copilot

Engineering teams at companies including GitHub, SAP, Barclays, American Express, EY and Unity use Codility.

AMERICAN EXPRESS

Codility is rated 4.6 out of 5 on G2 and ranked #1 for Enterprise Technical Skills Screening. Rankings come from verified customer reviews.

Today’s failing approach: ban AI, allow it without a record, or rely on self-reports

Challenge 1: AI is banned in the assessment

If your engineers build with AI every day, an assessment without AI does not reflect the way your team works.

Challenge 2: AI is allowed without a record

When your take-home process captures only the submitted code, you cannot see how the candidate used AI. The final code alone does not show whether the candidate understands the AI’s suggestions.

Challenge 3: AI skills on your current team are self-reported

If you rely only on manager opinion and self-assessment surveys, you have no task results to compare across your team. Those opinions alone do not show which engineers can build with a coding agent and check the result.

One platform to hire and verify AI-native engineers

Codility Screen assesses candidates before the live interview. Codility Interview lets your team question finalists as they build. Skills Intelligence assesses your current engineers. Candidates and engineers can use AI tools, including Claude Code, in each product’s VS Code environment.

Screen candidates on AI-native tasks

Codility Screen is an assessment each candidate completes on their own time. You choose an AI-native task from the Codility task library for candidates applying for the same role. An AI-native task is a real engineering task built around work with AI.

Work with AI tasks ask candidates to solve engineering problems with AI assistance. Some tasks include common AI suggestions that lead to an incorrect solution.
Scoring follows fixed rules against test cases mapped to the Engineering Skills Model, Codility’s public model of engineering skills. As a result, the same output always gets the same score.
Codility captures every prompt, acceptance, modification and rejection as reviewable AI activity in the candidate report.

The difference: You can review the task result and the candidate’s AI activity before deciding who advances.

ScreenBackend engineer, AI-nativePriya Raman

PR

Priya Raman Completed Oct 14, 2026 · 58 min

Completed Move to Interview

Summary Task details AI activity Code playback

82% Task score

Order export API · python

Skills

AI output validation92%

Critical thinking85%

Correctness78%

Testing74%

14Prompts

6Accepted

5Modified

3Rejected

TimeAI activityAction

04:12PromptThe last page of orders is missing from the export…Prompt

04:31Suggestion · line 13pages = total // size + 1Rejected

06:05PromptUse ceil and add a test for total=500, size=100Prompt

06:22Suggestion · test_export.pyCandidate added a size=1 boundary caseModified

09:48Suggestion · line 9size: int = Query(100, ge=1, le=500)Accepted

Live Final interview

AI Assistant Off AI Copilot Off Claude Code

PR JK

client.py test_client.py

181920212223

                      for attempt in range(self.max_retries + 1):
                          try:
                              return self._post(“/charges”, payload)
                          except (ServerError, Timeout):
                              time.sleep(0.5 * 2 ** attempt)
                      raise PaymentUnavailable()
                

Terminal · claude
priya@payments$ claude> Add retries with exponential backoff to PaymentClient.charge.  Don’t retry on 4xx responses. ● Read src/payments/client.py● Edit src/payments/client.py +18 −3● Bash pytest tests/test_client.py -q  ⎿ 6 passed in 0.41s charge() now retries up to 3 times on 5xx and timeouts,with 0.5s, 1s and 2s delays. 4xx errors raise immediately. > 

Live AI activity

Interviewer view

Prompt14:02Add retries with exponential backoff to PaymentClient.charge. Don’t retry on 4xx responses.

Response14:03Edited client.py (+18 −3). Ran 6 tests, all passed.

Prompt14:05Make max_retries configurable from settings.

Interviewer notes

Ask: why no jitter on the backoff delays?

Watch finalists build with Claude Code in a live interview

Codility Interview is the second stage, a live technical interview in a shared VS Code workspace. Your team invites the candidates it shortlisted in Screen. Interview also runs on its own, without Screen.

Your team enables or disables the AI Assistant, the AI Copilot and Claude Code per assessment. Your team can run one interview with AI off and a later interview with AI on.
Claude Code runs in the workspace terminal when your team turns the agent on.
Your interviewer can see each prompt and each AI response as it happens.
Your interviewer can then ask the candidate to explain the design decisions behind the code the AI wrote.

The difference: Your interviewers see how each finalist directs and corrects the AI before the final decision.

Assess AI skills across your current engineering team

Codility Skills Intelligence assesses the engineers you already employ using the same tasks and scoring. You can compare new hires and your current team on the same basis. Skills Intelligence runs on its own, without Screen or Interview.

Skills Intelligence gives each engineer a proficiency level per skill, including AI skills such as checking AI output.
Engineering leaders can compare teams side by side and see which teams have gaps in AI skills.
Your team can assess the same engineers again after AI training and see whether proficiency changed.

The difference: You plan AI training and staffing for your current team on assessed proficiency.

Request a free trial

Team comparison Engineering · 46 engineers assessed

Skill category: AI Round: Sep 2026

Skill Payments14 engineers Platform12 engineers Mobile11 engineers Data9 engineers AI output validation Advanced Proficient Gap Developing Working with coding agents Proficient Advanced Developing Gap Critical thinking Advanced Proficient Proficient Developing Code review Proficient Advanced Developing Proficient

2 teams with AI skill gapsMobile, Data

Re-assessment scheduledMobile · after AI training, Dec 2026

Questions about AI-native engineers

An AI-native engineer is a software engineer who builds with AI tools and coding agents as part of daily work. An AI-native engineer gives the AI context, breaks the work into steps, checks the AI’s output and fixes the AI’s mistakes. Teams also use the terms AI-native developer and AI-first engineer.

Hire and assess AI-native engineers with Codility

The walkthrough shows an AI-native task from the candidate’s side and the candidate report your team reviews afterward.

Request a free trial

"*" indicates required fields

Comments

This field is for validation purposes and should be left unchanged.

Business email*

First name*

Last name*

Job title*

Company*

Direct phone number

What can we help with? (select all that apply)*

Assess candidates (Screen)

Improve technical interviews (Interview)

Understand my team’s skills (Skills Intelligence)

Replace or evaluate your current platform

I’m an existing customer

Something else

Assess AI skills in business roles (AI for Business)

Please tell us more*

Roughly how many hires do you expect this year?*

How many employees does your company have?*

Anything you'd like us to know? (optional)

How did you find us? (optional)

Please tell us how you found us*

utm_source

utm_medium

utm_campaign

utm_content

utm_term

landing_page

referring_page

First touch date

First touch source

First touch medium

First touch campaign

First touch content

First touch term

First touch click ID

First touch referrer

First touch landing page

Most recent touch date

Most recent touch source

Most recent touch medium

Most recent touch campaign

Most recent touch content

Most recent touch term

Most recent touch click ID

Most recent touch referrer

Most recent touch landing page

GCLID

AB variant

Hire AI-native engineers who catch AI errors_

Today’s failing approach: ban AI, allow it without a record, or rely on self-reports

Challenge 1: AI is banned in the assessment

Challenge 2: AI is allowed without a record

Challenge 3: AI skills on your current team are self-reported

One platform to hire and verify AI-native engineers

Screen candidates on AI-native tasks

Watch finalists build with Claude Code in a live interview

Assess AI skills across your current engineering team

Questions about AI-native engineers

What is an AI-native engineer?

How do you hire AI-native engineers?

Which coding assessment platform has AI-native tasks for agentic engineering roles?

How do you test whether an engineer can build with AI?

Does Codility use AI to score engineers?

Hire and assess AI-native engineers with Codility_

Hire AI-native engineers who catch AI errors

Hire and assess AI-native engineers with Codility