Liping Zhang

About Liping Zhang

I am a Senior Staff Machine Learning leader specializing in search, ranking, and recommendation systems, with deep expertise in multi-task learning, multimodal AI, LLMs, VLLMs, and large-scale production systems. I am currently a Staff Lead ML Scientist at Adobe, leading the development of the Relevance AI Foundation powering Adobe Express and Stock. I drive end-to-end innovation from model architecture to production deployment and A/B impact, building scalable ranking systems that improve user engagement and business outcomes. At Adobe, I have delivered measurable impact, including +XX% CTR lift,+X% CVR improvement, and –XX% null rate reduction, while establishing reusable AI foundations across teams. I designed MTL/MMoE and Transformer-based ranking architectures, developed LLM-augmented and multimodal retrieval systems, and built automation pipelines that improved ML iteration speed by ~90%, bridging research and production systems. Previously, at Pinterest, I led the development of e-commerce search and recommendation systems across multiple surfaces, building deep learning models including two-tower recommenders, multi-task models, and personalized ranking systems that drove significant GMV and checkout improvements. Before that, at JD.com, I led development of large-scale ML and NLP systems for search and recommendation, contributing to distributed systems including HBase, Zookeeper, and Spark. I bring 16+ years of experience in large-scale search, retrieval, ranking, recommendation, and ads systems, with 14+ years of technical leadership and people management experience. I have authored 9 top-tier publications and filed 7 patents, advancing ranking, multimodal retrieval, and personalization. My focus is on building scalable AI systems, driving technical leadership, and delivering measurable business impact at scale.

Experience

1. Staff Lead Machine Learning Scientist, Sensei & Search

   Adobe

   Apr 2024 — Present · San Jose, CA, US

   Senior Technical Leader(Cross-Organizations TL) for Adobe Sensei and Search, Discovery, Content Intelligence, Recommendation, and Search ranking.• Led end-to-end development of ML ranking systems (V1–V7+), delivering strong product impact:+XX% CTR,+X% export rate, –XX% null rate reduction• Designed and productionized L3 multi-objective reranking systems (MTL/MMoE + Transformer, V1-V5+), unifying click, conversion, and relevance optimization• Built the Relevance AI Foundation, a scalable architecture supporting ranking and recommendation across Express, Stock, and future Adobe AI products• Architected ML automation pipelines, reducing iteration cycles from ~2 weeks to 1 day (~90% improvement)• Developed Transformer, LLM, and multimodal ranking systems (TCR / VTCR) with strong offline performance (~94% AUC,~88% NDCG@5)• Designed Bi-Encoder retrieval systems, improving recall and enabling scalable ranking integration• Led cross-org execution and mentored 15+ engineers, enabling independent delivery of ranking systems and A/B experiments• Authored 9 top-tier publications and filed 7 patents, advancing ranking, multimodal retrieval, and personalization

Education

• Hubei Enan Senior High School

  High School Diploma, Engineering

• Zhejiang University

  Master Degree, College of Computer Science and Technology, Computer Application Technology

• Hunan University

  Bachelor's degree, College of Computer Science and Technology, Computer Science and Technology

Skills

• Hadoop
• Java
• Hbase
• Linux
• Zookeeper
• Software Development
• Software Engineering
• Computer Science
• Distributed Systems
• Apache Open Source
• Pheonix
• Spark
• Solr
• Search
• Flume
• Hive
• Mapreduce
• Yarn
• Morphline
• Cloudera Edh
• Mahout
• Sql
• Scjp
• Pmp
• Data Mining
• Machine Learning
• Cloud Computing
• Text Analytics
• Big Data
• Cassandra
• Impala
• Scala
• Pig
• Bash Shell
• Python
• Biginsights
• Intel Idh
• Kafka
• Akka
• Nlp