Responsible for minimizing risks implied by inauthentic activities (bot/spam/fraud/abuse) on TikTok.The BRIC org I support is a global team consisted of product, operation, engineering, machine learning and data science experts. We are actively hiring all levels of talents in Singapore, San Jose and Seattle.

@ Site Integrity: Supported 3 teams: foundation (ML infra), fake engagement (inauthentic social graph edges), spam (inauthentic content distribution).@ People Data Platform: Supported the workforce that built social graph user clusters wrt to their privacy preferences.

@ Product Safety: Manager the anti-spam & anti-abuse teams.Led Twitter's engineering effort against spam & abuse. Explored a portfolio of product initiatives. Drove cross organization infrastructure innovations. Increased operational efficiency by bettering visualization & investigation tools. Applied a combination of policy driven / heuristic rule based / supervised & unsupervised ML algorithms to automate majority of detection.

@ Product Safety: Tech lead of the anti-spam team.Drove the development & application of Twitter's real-time rule engine for safety, with its extension in visualization, stream computing, ML classification & clustering.

@ Value Based Bidding (VBB) team in Google Display Network (GDN) org.Using post-ad-click conversion data to calibrate future real-time CPC bids to given ROI target. Products involved included Conversion Optimizer, Enhanced CPC, Target CPA and Target Return-On-Ad-Spend (ROAS).Led and developed launching of Target ROAS in 2013, allowing advertisers to directly control their actual dollar-cost per dollar-revenue.

@ Shifen, an advertiser-facing engineering team.Ads keyword recommendation. Profit analysis across ad products.

@ Digital Intelligent Tools Group: Computer Vision -- pedestrian detection & image classification.Co-authored a state-of-art work published in NIPS'2008.