Nir David enhanced the calibration workflow in the vllm-project/vllm-gaudi repository by developing a Calibration Configuration Generation Enhancement for calibrate_model.sh scripts. Using bash scripting and automation skills, Nir introduced a calibration sample interval parameter to the measurement configuration files, improving model calibration accuracy and consistency. The solution addressed configuration drift by ensuring cross-script consistency, enabling repeatable calibration runs across different environments. Nir collaborated closely with team members to validate and document the changes, consolidating the fix across multiple scripts. The work demonstrated a focused approach to DevOps practices, emphasizing reliability and maintainability in the calibration process without introducing unnecessary complexity.