● Architected and implemented an AWS-based Intelligent Document Processing System that achieved 70% reduction in document processing time.● Led the development of a multi-country HS code classification system using LangGraph and GPT-4o, resulting in coverage expansion to 5+ countries.● Designed and deployed a collaborative data annotation platform using that reduced team training time by 30%.● Fine-tuned a language model for Named Entity Recognition (NER) to automatically extract key fields from customs import/export invoices, achieving 95% extraction accuracy.● Implemented AI-powered HS code classification system for customs declarations, enabling real-time product categorization with 90% accuracy across 420+ categories.

● Managed and performed data collection and labeling of more than 100k+ examples for a text classification model with a team of 5 members.● Implemented automated data preprocessing pipelines using Python, reducing data preparation time by 50%.

● Worked on 6+ international projects for clients in the pharmaceuticals, advertising, and academic industries.● Implemented ETL pipelines integrating Business Central, Xero, and SQL databases with automated daily refreshes, achieving 99.9% data accuracy and reducing manual processing time by 60%.● Developed comprehensive monthly and weekly dashboards in Zoho Analytics for an Australian service-based company to visualize numerous KPIs including sales and growth.