Coding assessment engine
The coding interview that actually proves they can code.
A real editor, real test cases, ten real languages. No HackerRank integration required, no extra invoice, no separate login built into every Paraakh exam and assessment pipeline.
Monaco editor 10 languages Hidden tests Real-time exec
solution.py tests.txt PYTHON
1 def reverse_list(arr):
2 return arr[::-1]
3
4 # run sample tests
5 print(reverse_list([1, 2, 3])) Test cases 4 / 5 passing
Sample 1: [1,2,3] [3,2,1] 12 ms
Sample 2: [] [] 8 ms
Hidden 1: [5,5,5] [5,5,5] 11 ms
Hidden 2: [1..10⁴] [10000..1] 84 ms
Hidden 3: [1..10⁶] Time limit exceeded TLE
Definition
A coding assessment is a hands-on test that measures a candidate's real programming ability they write and run actual code against hidden test cases, instead of answering multiple-choice questions about code. Paraakh runs it in the browser on a VS Code-grade editor, auto-grades every submission on correctness and efficiency, and drops the result straight into your hiring or exam pipeline.
Ten languages, one engine
Test what they'll actually build in.
Same editor, same test runner, same scoring across every language your team writes.
What's in the engine
Eight things that make a coding screen defensible.
Monaco editor
The same editor that powers VS Code syntax highlighting, autocomplete, multi-cursor, find-and-replace. Candidates get a real editor, not a textarea.
Hidden test cases
Candidates can't see the expected outputs that decide their score. No gaming by reverse-engineering the assertions.
Open sample tests
Per-question sample tests the candidate can run to validate their approach before submitting.
Custom input runs
A scratch input box for debugging paste in any input, run, see the output. Encourages real engineering hygiene.
Function-based challenges
Function metadata (name, params, return type) and idiomatic starter code per language no boilerplate-juggling to test the actual skill.
Per-test limits
Configurable time and memory limits per test case. Catch O(n²) solutions on the right data sets automatically.
Real-time feedback
Live execution progress via WebSocket each test case ticks green or red as it runs, with stdout, stderr, and runtime.
Submission history
Every run saved with timestamps. Candidates mark a submission as final, you see every attempt that led there.
Don't have a coding question bank?
Generate full coding problems with AI.
Starter code per language, sample tests for the candidate, hidden tests for scoring generated, validated, and ready to ship in seconds. Reviewed by you before any candidate sees them.
See AI question generation → AI-GENERATED
Two-pointer · medium
- Starter code · Python, Java, JS, C++
- Sample tests · 3 visible cases
- Hidden tests · 8 edge-case cases
- Time limit · 200 ms · 256 MB
- Quality score · 94 / 100
Where it fits
Built into every assessment, not bolted on.
Use as a round in a multi-stage pipeline Paraakh advances only the candidates who pass the coding cut-off.
See the hiring pipeline →A coding section inside an exam blueprint sectioned timing, mixed with MCQ, proctored end-to-end.
See secure exams →Every student writes real code under proctoring. Per-student question sets prevent paper leakage between batches.
See Paraakh Campus →Validate architecture, API, and security knowledge with hands-on coding before allocating someone to a client engagement.
See Comply → Real code, real signal
Pick a problem. Watch a candidate solve it live.
Bring a coding problem or let us generate one and we'll show you the editor, the test runner, and the result that lands in your shortlist. Eight minutes, real product.