Seamus W.

AI-powered document extraction, OCR and scraping is my core specialization. You are not paying me to experiment; you are paying for quick, reliable execution from an engineer who knows this solution space inside and out.

While I have a broad background in AI (Master's from King's College London, 5+ years LLM development), my niche is turning messy, unstructured data into clean, validated databases.

Why Work With Me?
Many developers try to force outdated techniques onto strict data extraction problems, leading to hallucinations, broken schemas, and costly architectural redos. I stay up to date on all modern data extraction literature. I know the pros and cons of every method, from modern vision models for implicit table relationships to custom OCR fallbacks.

Because I know the pitfalls of document extraction and web scraping before we even write the first line of code, I can tailor the exact right architecture for your solution from day one. No guessing. No structural redos. Just fast, production-ready execution.

Experience:
I have built production-grade extraction pipelines and automated scraping architectures across multiple complex domains:

• Financial & Invoicing: Engineered invoice extraction pipelines achieving 99%+ accuracy across highly varied layouts, mapping dates, amounts, VAT, and line items.

• Procurement & Tenders: Processed 200+ page Brazilian tender documents, identifying and extracting only specific supplier items and complex split tables from hundreds of pages of mixed content.

• Sports & Large-Scale Data: Currently building a centralized database of global triathlon results, requiring massive-scale web scraping and data extraction across thousands of poorly formatted, highly variable PDFs.

• Legal & Regulatory: Built an AI extraction pipeline using daily cron jobs to scrape government websites, monitor policy updates, version new documents, extract structured changes using AI, and alert affected users.

Beyond Extraction: Putting Your Data to Work:
Getting the data is only half the battle. I also build the systems that make your newly structured data actionable:

• CRM Integrations & Data Enrichment: Automatically routing validated extraction data (like invoice details or scraped leads) directly into Salesforce, HubSpot, or custom databases to enrich your existing records.