AI Benchmark Task Author – Full-Time Internship

Experience: Freshers / Students / Early-Career Developers

Type: Full-Time Internship

Location: Remote

Duration: 3–6 Months

Stipend: ₹35,000/month for the first 3 months

Working Days: Monday–Saturday | 8 Hours/Day

Role Overview

We are looking for technically strong interns to create challenging, real-world engineering tasks for AI coding agents . You will own tasks end-to-end—from designing the problem to developing the reference solution and building a reliable automated grader.

Key Responsibilities

- Design realistic engineering and coding challenges for AI agents.
- Write clear task requirements and technical specifications.
- Build correct reference solutions and automated verifiers/graders .
- Create rigorous tests covering edge cases and incorrect solutions .
- Work with Git/GitHub, Docker, Linux, and multi-container environments .
- Test, debug, and improve tasks until they meet required quality standards.

Key Requirements

- Strong programming skills in Python, C/C++, TypeScript, Verilog/RTL, or similar .
- Hands-on knowledge of Git, GitHub, Docker, and Linux .
- Experience writing automated tests .
- Strong debugging, analytical, and problem-solving skills.
- Familiarity with AI coding tools/agents .




- Good technical writing and ability to work independently.

Internship & Performance Details

- First 10 days: Closely monitored trial period.
- First month: Strict availability from 10 AM–7 PM IST .
- Performance target: Minimum 60 accepted tasks/month.
- Each task typically requires 4–5+ hours , depending on complexity and skill level.
- After successful trial and consistent performance, working hours can become adaptable while maintaining the monthly target.
- Compensation may be revised after 3 months based on performance.

Selection Process Application → 15-Minute MCQ Assessment → 1-Hour Build Challenge → Techno-Managerial Interview → Selection The assessment covers Git, Docker, Linux, Testing, Problem Solving, Python, and Task Design . Candidates must complete the assessment and build challenge within 24 hours of starting the application process .

The assessment and build challenge must be completed independently. Copy-pasted or AI-generated submissions will be rejected.

Apply Here

AI Benchmark Task Author Assessment

Shortlisted candidates will be contacted via email after review, which may take up to one week .

Queries: [email protected]

📌 AI Benchmark Task Authors (Internship) (India)
🏢 Codefeast
📍 India