joywekesa082-arch/Swahili-ai-data

examples of data labelling, translation quality evaluation and prompt engineering for Swahili AI dataset # Swahili-ai-data welcome!This repiratory showcase my cababilities in data annotation,linguistic quality assurance and RLHF( Reinforcement Learning from HumanFeedback) prompt evaluation for swahili Large Language Models (LLMs). ---- ## COre capabilities * **Data annotation& Tagging :*structuring unstructuring swahili text for classification models. * **Linguistic QA:*identifying subtlegrammatical, contextual, and cultural translation issues. * **Prompt Engineering & Evaluation:*Assessing LLM outputs for safety,helpuflness, and localization accuracy. ---- ## 1.Text classification& Intent Tagging * ** Objective:**Categorizing user queries to train conversational AI assistants. * **sample Data:** |User Query(swahili)|Primary Intent|Sub Category Intent|Language Quality Notes| |:---|:---|:---|:---| |"Ninasanduku gani la malipo na nawezaje kutoa pesa kwa m-pesa?"|Financial_Inquiry|withdrawal_method|standard colloquial syntax,includes regional terminology(M-pesa).|programu hii haifanyi kazi."|Technical_support|App_cash|High Agency tone;|