Open Train AI Capabilities — Finance & Tax data annotation, NER/classification, and RLHF/SFT evaluation (current contributor).
DocumentHandled large-scale labeling and quality evaluation of East African financial and tax documents to support downstream AI tasks like entity tagging. Applied expertise to identify inconsistencies across accounting records (e.g., invoice vs LPO vs GRN) and flag anomalies in amounts, dates, and tax codes. Contributed domain-specific guidance for RLHF/SFT/red teaming-style evaluation of AI outputs in finance and compliance contexts.• Labeled/validated entities such as vendor names, account numbers, amounts, and dates in business/tax text.• Classified document types relevant to finance workflows (e.g., invoice, receipt, contract) using accounting conventions.• Evaluated prompt/response accuracy and compliance for AI financial advice, including hallucination detection.• Supported multilingual annotation considerations for English (native) and Swahili (fluent) business documents.