AI Task Architect for Backend Evaluation

Mindrift; Doha; On-site; QAR 150,000 - 201,000; Full time

Job summary

Mindrift is seeking a seasoned backend-focused developer task designer. You will craft realistic coding challenges in Docker, assemble self-contained task packages, and write deterministic pytest-based tests to verify fixes. Deliverables include broken code, tests, instructions, and a reference solution.

This is project-based, not permanent employment. You will design tasks that resemble real-world developer work, ensuring reproducible environments, clear Jira-like tickets, and scalable

Qualifications

• 3+ years of production software development in one backend stack.
• Python + pytest fluency — required regardless of primary stack.
• Experience with Docker, Linux and Bash in containerized environments.

Responsibilities

• Invent realistic developer scenarios — real bugs, missing features, or broken ETL.
• Build reproducible Docker environments with pinned dependencies.
• Write a pytest-based verification that is deterministic and non-flaky.
• Create an instruction.md that reads like a Jira ticket for a developer.
• Provide a reference solve.sh proving the task is solvable.
• Calibrate difficulty so state‑of‑the‑art agents solve the task 20–60% of the time.
• Review other authors' tasks as a QA reviewer later.

Skills

Backend development

Python

pytest

Docker

Linux & Bash

AI tooling familiarity

Tools

Docker

pytest

CI/CD

Job description

Mindrift is seeking a seasoned backend-focused developer task designer. You will craft realistic coding challenges in Docker, assemble self-contained task packages, and write deterministic pytest-based tests to verify fixes. Deliverables include broken code, tests, instructions, and a reference solution.

This is project-based, not permanent employment. You will design tasks that resemble real-world developer work, ensuring reproducible environments, clear Jira-like tickets, and scalable