Ryan Wu Email & Phone Number

Mountain View, Ca, Us

Lead Manufacture AI Team within Google Hardware to empower# Machine Learning foundation of massive AI solutions from modeling, delivery and operations in Manufacturing, (e.g. AI powered visual inspection)# Quality Operation through advanced AI capability such as anomaly detection and predictive maintenance# Advanced NLP platform that analyze customer data and produce actionable insights for Voice of Customers

Beijing, Cn

* Built Relevance Organization with 20+ members (includes Infrastructure/Modeling/Analytics) from Ground-Up* Cultivated/Enabled Relevance Awareness across various functions (BUs) for collaborative work on # Core Product Relevance: 1. Designed & Rolled-out reinforcement learning based decentralized recommendation algorithm that significantly uplift user engagement, retention and eventually DAUs 2. Enabled relevance components with core product innovations # Monetization Relevance: Enabled relevance arms in MM dollar new product lines and personalized pricing strategies # Anti-Spamming: Improves spammer detection through deep image processing and delivers relevance models that blocks inappropriate photos & conversations for promoting better authentic/clean communityPaper in Preprint: Scientific Experimentation and Reciprocal Recommendation: A Tantan Practice

Sunnyvale, Ca, Us

* Lead recruiter relevance team to deliver product and continuous relevance improvements for LinkedIn Recruiter # Deliver interactive recommendations for Recruiters to optimize search attributes selection process # Drive on search relevance/ranker optimization to deliver innovative Candidate Profile based retrieval # Drive continuous relevance innovation for improving hiring efficiency/effectiveness through Machine Learning & Deep Learning models # Granted Patents -- US-10373075-B2: Smart Suggestions for Query Refinements -- US-10409830-B2: System for Facet Expansion -- US-10984385-B2: Query Building for Search by Ideal Candidates -- US-10606847-B2: Generation of training data for Ideal Candidate Search Ranking -- US-10860670-B2: Factored Model for Search Results and Communications Based on Search Results -- US-10769136-B2: Generalized Linear Mixed Models for Improving Recruiter Search -- US-10482137-B2: Nonlinear Models for Member Searching -- US-10726025-B2: Standardized Entity Representation Learning for Smart Suggestions -- US-11436522-B2: Joint Representation Learning of Standardized Entities and Queries -- US-10956515-B2: Smart Suggestions Personalization with GLMix -- US-10628432-B2: Personalized Deep Models for Smart Suggestions Ranking Publications:WWW 2016: Search by Ideal Candidates: Next Generation of Talent Search at LinkedInCIKM 2017: From Query-By-Keyword to Query-By-Example: LinkedIn Talent Search ApproachSIGIR 2018: Talent Search and Recommendation Systems at LinkedIn: Practical Challenges and Lessons Learned

Us

Driven the back-end data strategy and product development in scalable data acquisition, data presentation to data intelligenceBeing the core contributor of full data technology stack (70% out of the whole code base) * Architected the distributed sales focused crawler with adapted Machine Learning based web scaper to robust scrape internet for sales intelligence in efficiency and at scale millions of pages * Architected full line of text processing from NLP processing (entity recognition), de-duplication to clustering, and categorization * Architected and built the data back-end utilizing search technologies as Lucene, big-data technologies such as Redis, Kafka, Storm to support business growth at scale and in real-timeFull stack hands-on entrepreneurship * Offshore R&D office set-up and team building from zero to disparate functional teams (~50) * Internal software process Definition and organizational IT control * Production Operation and Failure Monitoring Skills involved Distributed computing: Hadoop/Storm, Redis/Kestrel/Kafka, Avro RPC Machine Learning: classifications as SVM/Logistic Regression, clustering with Locality Sensitivity Hashing/Min-Hash, Pig/R/Scikit-Learn, Text learning as Vowpal Rabbit, StanfordNLP NoSQL/MySQL databases Search Technologies: Lucene / SOLR

Redwood City, California, Us

Developed software backbone for analysis and data servingDeveloped data processing pipeline for anomaly detection

Designed/Implemented LDA based user profiling algorithms over Hadoop/HBase clustersDesigned/Implemented nearest neighbor clustering and collaborative recommendation algorithms

Ph.D, Electrical Engineering

Master'S Degree

Bachelor’S Degree, Communication Engineering