Naod Girma

Performed RLHF and SFT evaluations on LLMs, rating prompts and responses for accuracy, safety, and instruction adherence. Acted as SME for medical and STEM content, identifying hallucinations and inaccuracies with high quality accuracy.

Evaluated and optimized multilingual generative AI datasets for English and Amharic, ensuring cultural and linguistic accuracy. Applied MQM frameworks for translation quality assessment and audited model responses for safety and policy compliance.

Managed clinical documentation and patient care workflows for 15-20 patients daily. Served as Principal Investigator for a research project on Venous Thromboembolism, including study design, data collection, and statistical analysis.

Executed high-precision video annotation, object tracking, and segmentation for computer vision model training. Maintained high accuracy through systematic cross-verification and resolved edge cases.

Pursued Doctor of Medicine (MD) degree, including clinical training and research, with a focus on medical practice and patient care.