Nazmus S.

AI can write a scraper in minutes. Keeping it alive through anti-bot, logins and site changes is the hard part, and it's what I do: ~12M posts/day in production, 547 sites in one pipeline, AI agent work at renowned AI startups. If your product, AI agent or data business runs on web data every day (not a one-time CSV), it's only as good as the data going in. I own that layer end to end: web scraping, data pipelines/ETL and AI browser automation. I build it, deploy it, monitor it and keep it running. "He did what another 15 developers tried to do but failed." (Upwork client) AI-ERA EXPERIENCE • AI Agent Engineer at a leading AI browser-agent startup: built AI-agent browser automation for enterprise customers, from requirements to deployed pipeline. • QA Manager at a well-known startup whose AI agents write scrapers: promoted from QA Engineer. Reviewed 20–100 AI-generated scrapers a day and catalogued where they break: dynamic pagination, JS-rendered content, stateful sessions. • Self-healing multi-agent system for government documents: agents categorize sites, scrape, grade the output and auto-repair failures, with no human in the loop. • LLM-powered expired-domain evaluator: estimates value from historical content, backlinks and niche. PROOF AT SCALE → Social media analytics platform: built from scratch and run a YouTube scraping system processing ~12M posts/day on GCP/Kubernetes. Caching + scheduling cut cloud costs ~40% with no loss of data freshness. → 547 websites, one config-driven pipeline on AWS: auto-scales, isolates failing sites, emails reports, zero manual steps after deploy. → Facebook scraping at scale through aggressive anti-bot, after 2 other developers had failed. Industries: lead gen & sales intelligence, e-commerce (eBay, Amazon), real estate, public records, social media data, academic research (Harvard University). Top Rated · 100% Job Success · 4.99/5 from 41 reviews · on Upwork since 2019 "Nazmus has become my go-to web automation developer for all my scraping and app-building needs in the ETL process." HOW I WORK ✓ Sample data or a proof of concept before you commit ✓ Monitoring, and I stay reachable after delivery ✓ 4+ hours daily overlap with US Eastern time; happy to get on a call ✓ Best fit: recurring data feeds and scrapers that need an owner. Not a fit for data entry or manual list building. Stack: Python, Scrapy, Playwright, Selenium, FastAPI, PostgreSQL, MongoDB, AWS, GCP, Docker, Kubernetes, Terraform