iamhun

Built the court data scraping pipeline from scratch, collecting jurisprudence records from multiple Brazilian tribunals (STJ, TJs) using Node.js and Puppeteer, handling anti-bot measures, pagination, and document extraction at scale. Processed and normalized raw legal documents into structured formats, implementing data validation and transformation logic to ensure consistency across tribunal sources. Integrated scraped data into a semantic search pipeline, preparing and indexing legal content for AI-powered retrieval — enabling the platform to match natural-language queries to relevant court decisions rather than relying on keyword search.