Michal B.

Tell me what data you need and where it should end up. I'll build the system that collects it, checks it and keeps it coming. That's what I've done for 10 years. I get data out of sites that block bots, read numbers out of PDFs and scans, and connect APIs that don't talk to each other. Now AI is part of every build. The work goes faster, and every value still keeps its source so you can check it. Work I take on: • Data from hard sites (Cloudflare, Akamai, DataDome), collected on a schedule and kept running for months • Research and list building done by automation, with a source link on every row • PDFs, scans and filings turned into a spreadsheet or database, with the file and page for every value • Syncs between Shopify, Zoho, Supabase, HubSpot, Google Sheets and your own database • Finishing Python tools and AI prototypes so they run on their own, with logs and alerts • Written reviews of an existing database, script or pipeline, with a list of what to fix first • MCP servers that let Claude or ChatGPT use your data and tools Whatever the source, records get matched and deduplicated, so each company or person shows up once and a rerun never adds duplicates. Some things I've built: a due diligence platform that cross-checks 14 UK regulators and professional bodies and flags issues for every firm. Scrapers for 5 US state procurement portals that kept blocking the previous team. A news-to-trade system where LLMs read every article and 15+ checks decide what gets traded, running on real brokerage accounts. A football stats feed into Google Sheets that has run on its own since 2023. Fixed price is fine. Send me what you have and I'll tell you how I'd do it.