Scientific AI Evaluation, Computational Problem Designer

This is a fully remote position, open to applicants in United States.

📋 Description

• Develop complex computational challenges that utilize specialized scientific software.

• Design tasks that assess both accurate execution (multi-step workflows, simulations) and strategic thinking (experiment planning, inference from incomplete data).

• Construct problem frameworks, solution strategies, and validation processes.

• Adjust and enhance tasks based on model performance to reach desired difficulty levels.

• Ensure that challenges prioritize reasoning strategies over mere brute-force computation.

⛳️ Requirements

• Proven expertise with at least one pertinent scientific library (through research, open-source contributions, or industry experience).

• Capability to work autonomously and adapt based on feedback.

• Proficiency in Linux/terminal environments and familiarity with remote computing configurations.

• Commitment of at least 15–20 hours per week.

• Advanced academic credentials (MS or PhD preferred) in a relevant STEM discipline.

• Practical experience utilizing scientific software libraries for genuine research challenges.

• Strong skills in Python programming, including the development of computational workflows and validation tools.

• Ability to craft intricate problems that necessitate deep reasoning rather than superficial solutions.

• Understanding of edge cases, constraints, and real-world challenges associated with scientific tools.

• Experience spanning multiple domains or tools.

• Knowledge of evaluation frameworks or benchmarking methods.

• Background in teaching, pedagogy, or design of problem sets.

• Awareness of reproducible research practices and containerized environments.

🏝️ Benefits

• [Insert specific benefits here]

• [Insert additional benefits here]