Foundation Models Researcher - Calibration & Truth

Flashmind Labs; Dubai; On-site; AED 420,000 - 660,000; Full time

Job summary

Flashmind Labs in Dubai is seeking researchers to push the boundaries of foundation models optimized for truth. You will contribute to novel training objectives, uncertainty-aware architectures, and evaluation methods, collaborating with scientists and engineers.

The role focuses on empirical outcomes, calibration, and robustness across distribution shifts, requiring strong independent work and a track record in advanced research.

Qualifications

• PhD or equivalent in ML/CS or related field.
• Proficiency in Python and PyTorch.
• Ability to design, run, and analyze experiments independently.
• Understanding of large-scale training and accelerator-based compute environments.

Responsibilities

• Develop novel training objectives for factual accuracy and calibration.
• Design architectures that handle uncertainty and distributional information.
• Build training pipelines grounded in empirical outcomes.
• Develop evaluation frameworks for calibration, tail event accuracy, and robustness under distribution shift.

Education

PhD or equivalent in ML/CS or related field

Tools

Python

PyTorch

Job description

Flashmind Labs in Dubai is seeking researchers to push the boundaries of foundation models optimized for truth. You will contribute to novel training objectives, uncertainty-aware architectures, and evaluation methods, collaborating with scientists and engineers.

The role focuses on empirical outcomes, calibration, and robustness across distribution shifts, requiring strong independent work and a track record in advanced research.