Invisible Technologies — AI Language Trainer / Language Specialist
Provided language scenario testing for AI model responses by conversing with AI models and evaluating output quality. Verified factual accuracy and logical soundness, then documented errors and proposed improvements to prompt engineering, training methods, and evaluation metrics. Actively forced mistakes to improve AI safety and accuracy while capturing reproducible error traces for the team. • Evaluated Malayalam communication and understanding performance • Performed error tracing and reported findings for iterative improvements • Reviewed response quality using factual and reasoning checks • Contributed to training/evaluation metric refinement