Skip to main content

Challenges Leaderboard Jobs Blog Pricing Hiring?

// Challenge library

Challenges

Complete a challenge to add a verified artifact to your public profile. Build in any language — evaluated on behaviour, not syntax.

// not ready for a full challenge?

~10 min

Try a prompt challenge — write a system prompt in the browser, scored in seconds. No setup needed.

Meeting Action Items →RAG Relevance Judge →

RAG Pipelines AI Agents MCP Servers Coding Agents Evals & Testing AI Tool Proficiency

Difficultyeasy medium hard

// 5 full challenges · Evals & Testing

Evals & Testing

Write evals for prompt regression testing

Design a lightweight eval suite that catches regressions in a structured extraction prompt.

Avg 2–3 h·Any language

Evals & Testing

Write evals for a summarisation model

Design and implement a rigorous evaluation suite for an AI summarisation system. Your evals should catch common failure modes.

CTryCrucible

AI challenge platform by webpipl. Prove your skills with real code.

Platform

Challenges Leaderboard Jobs Blog Sign in

Companies

Why TryCrucible Pricing ATS webhooks Get access

WebPipl

webpipl.com Reviewers Contact us Privacy policy Terms of service

© 2026 webpipl. All rights reserved.

Avg 8–12 h·Any language

View →

Evals & Testing

Write evals for prompt injection defense

Design evals that detect prompt injection attempts in retrieved context and verify safe behavior from a RAG or tool-using assistant.

Avg 8–12 h·Any language

Evals & Testing

Write evals for RAG answer faithfulness

Design evals that detect unsupported, contradicted, and partially supported answers from a RAG system.

Avg 8–12 h·Any language

Evals & Testing

Write evals for a text sentiment classifier

Design an evaluation suite for a sentiment classifier (positive/negative/neutral) that surfaces known failure modes: sarcasm, mixed sentiment, and domain-specific language.

Avg 4–6 h·Any language·1 engineer completed

// scenario challenges

Scenario Challenges

One system prompt. Three different test inputs. Scored instantly in the browser — no GitHub repo, no local setup. About 1 hour. A good step up from a 10-minute prompt challenge.

Evals & Testing

LLM Output Judge — 3 Task Types

Evaluate LLM outputs for correctness, completeness, and quality across 3 different tasks

~1 hour·In-browser·3 scenarios