Get to know developers through a real project
Candidates work in their own editor and terminal, giving your team code, commit history, and design decisions to talk through in the interview.
4.9
4.9
5
I like how the library challenges are structured around on-the-job skills. The experience for candidates is excellent. They work locally with the IDE and tools they are most comfortable with.
Kevin Sahin
Co-Founder @ ScrapingBee
Built locally, reviewed like a pull request
Candidates code in their normal development setup, while your team reviews their changes, commit history, and AI review notes in one place.
Familiar local setup
Candidates write and debug code using their preferred editor, shortcuts, local packages, and terminal tools.
Reviewable diffs
Inspect commit history, file changes, and module structure across the project just as you would in a teammate's pull request.
Plan follow-up questions
Use the AI review notes to find areas worth discussing. Read the suggested questions and choose which files and tradeoffs to cover with the candidate.
Take-home challenge
Frontend Engineer Challenge
Terminal
$ git clone git.codesubmit.io/acme/frontend-challenge
$ cd frontend-challenge && npm install
$ git push origin main
→ Solution submitted
AI review ready
12 files reviewed · 4 follow-up prompts
Submission checks & handoffATS
Warden submission check
Passed
AI review
12 files · 4 prompts
Hiring stage: Senior Engineer pipeline
Project setup and tooling
Choose a project that fits the role
Start with a library project in your stack, draft a custom assignment from a job description with AI, or bring your team's own starter repository and tests.
Realistic project tasks
Assign multi-file features, bug fixes, and API integrations modeled on everyday engineering work, or import your team's own starter template.
Dedicated candidate repos
Each candidate works in an isolated Git repository, preserving branch history, commit messages, and diffs across every changed file.
AI review summaries
Get an overview of how the submission is structured, where test coverage is thin, and which questions are worth asking on the call.
60+ languages and frameworks
Assess across backend, frontend, mobile, and systems stacks, from TypeScript, React, and Python to Go, Rust, and Ruby on Rails.
Automated test suites
Include unit and integration tests in the starter project so candidates can check expected behavior locally and reviewers can inspect passing and failing cases.
Challenge remixing with AI
Paste a job description, specify the role level, or start from a library project to draft a custom assignment your team can edit.
CodePair continuation
Open a candidate's submitted project in CodePair when you want to walk through their implementation and pair on a follow-up change.
Build on their submission in CodePair
Once your team has read the submission and AI review notes, open the project in CodePair to talk through tradeoffs, add a requirement, and see how the candidate works with shared AI chat.
Extend their existing code
Start from the files and tests the candidate already wrote, then introduce a new edge case or feature change to see how they adapt their design.
Shared AI chat in the session
Both of you see the AI chat in CodePair. Follow what the candidate asks, inspect the suggested code together, and discuss what they keep or change.
Clear notes for the debrief
Bring the code changes, suggested review questions, and observations from the CodePair call to your hiring discussion, with concrete examples from the work.
CodePair AI Follow-Up
AI readiness review
When you continue a take-home submission in CodePair, this view summarizes how the candidate uses shared AI chat while working through follow-up tasks.
Illustrative scores
Shared AI chat
Optional code evidence
Critical Evaluation
8.2
Asks for tests, validation steps, or tradeoff explanations instead of accepting the first response.
Prompt Quality
8.0
Specific asks with relevant repo context, constraints, and acceptance criteria.
Task Ownership
7.8
Uses AI for bounded steps while directing the design.
Iterative Troubleshooting
7.6
Follows up, narrows scope, and builds on previous output when the first answer is not enough.
Sample reviewer output
Strong AI judgment
Score
79
/ 100
Candidate supplied relevant code, narrowed a request when the first reply was too broad, and asked for tests and tradeoff explanations before choosing an approach.
Evaluated during CodePair follow-up sessions created from take-home submissions
These four dimensions help reviewers examine how a candidate frames prompts, checks suggestions, directs the task, and iterates in the shared AI chat.
Adds code evidence when available
Architecture ownership
Debugging independence
Integration quality
When session diffs, terminal output, or test runs are available, reviewers can also inspect how generated suggestions were edited and integrated into the project.