1
1
China
1 year of experience
AI builder focused on evidence-grade web data agents, market transparency, and reproducible audit workflows. I build tools that combine live public web access, structured extraction, visual evidence, and careful claim boundaries so AI systems can turn messy web pages into trustworthy decision support.

ShadowBuyer is a Booking market transparency evidence agent powered by Bright Data Web Unlocker. It turns country-targeted public web access into replayable audit packages with raw HTML captures, rendered screenshots, SHA-256 hashes, and bounded review outputs. The main evidence layer is an SG/US repeated replay package: 20/20 clean captures, repeated sampling, fingerprint control, disclosed jitter, and a condition gate that separates price evidence from taxes-and-fees presentation differences. This lets the agent preserve what was actually observed instead of flattening everything into disposable scrape rows. A separate four-country live proof demonstrates single-sample live expansion across US, SG, GB, and DE. It is presented only as an expansion proof, not as a statistical conclusion. ShadowBuyer is not a generic crawler, not a full real-time monitoring SaaS, and not a legal discrimination claim. It is a bounded consumer-facing transparency audit tool designed to show how live public web data can become reproducible, reviewable evidence.
31 May 2026