At Turbolab Technologies (ScrapeHero), I lead the team behind custom scraper APIs, scheduled spiders, and RPA automations. I cut a location-based scraper API's average response time from 18–24 seconds to 5–8 seconds by caching store setup in Redis and running independent requests concurrently.
I co-built Spider Healer, which diagnoses broken scrapers and opens Draft merge requests for review; in testing, it fixed 8 of 10 cases and correctly stopped on the other 2. In R&D, I developed anti-bot configurations that reached ~80% success, and I designed an extraction pipeline that returns validated JSON at ~95% accuracy.
