Jz Lu

About Jz Lu

8+ years full-time experience of building end-to-end Data Science/ML/Deep Learning solutions including: user action prediction (ctr, cvr, etc), customer behavior modeling (LTV, profiling & embedding, segmentation, etc), recommendation systems, attribution modeling (multi-touch attribution), causal inference (A/B Testing, market mixed modeling etc), AdTech, and time series forecasting.* 3+ years experience of leading data engineer/scientist to deliver ML-powered products.* Hands-on experience building end-to-end AI applications, including fine-tuning LLMs, retrieval-augmented generation (RAG), and multi-agent frameworks.* Proficiency in Python, Spark, SQL, AWS, and MLOps for building and deploying production-grade ML solutions.* Solid problem-solving, project management, and strategic-thinking ability;* Founder of opentimeseries.com; Author on Medium/Toward Data Science.

Experience

1. Sr Applied Scientist

   Amazon

   Dec 2021 — Present · Palo Alto, CA, US

   Ad performance optimization: Building models to maximize advertiser value/surplus through bidding, ranking, pacing and charging optimization. · Forecasting & Recommendation: Developed, and deployed scalable ML/deep learning models for user response prediction, ad performance forecasting, ad inventory forecasting for Amazon Advertising (ADSP - Amazon Demand Side Platform). · Casual inference: Conducted A/B tests and causal impact analyses (e.g, diff-in-diff, Bayesian inference) to quantify improvements in traffic/revenue and ad performance across different user segments.· MLOps: Maintained production-grade code for training/inference pipelines using Python, Docker, and AWS services. · LLM Applications & Multi-Agent Systems: Build an AI-diagnostics platform that integrates LLMs (Bedrock) with retrieval-augmented generation (RAG) and multi-agent coordination to deliver campaign diagnostics, forecasting insights, and recommendations. · Research: Published peer-reviewed research papers at internal Amazon Machine Learning Conferences (acceptance rate <10%).

Education

• Donghua University

  Bachelor of Science (BS), Textile Sciences and Engineering

  2009 — 2012

• North Carolina State University

  Doctor of Philosophy (Ph.D.), TTM, User Behavior Modeling and Prediction;

  2014 — 2017

• North Carolina State University

  Master’s Degree, Consumer Behavior Modeling and Prediction, Statistics;

  2012 — 2014

Skills

• Materials in Textile
• Pencil Sketch
• Apparel Sourcing
• Hadoop
• R
• Erp Software
• Adobe Photoshop
• Crm
• Supply Chain Optimization
• Sql Server Integration Services (Ssis)
• Consumer Behaviour
• Social Media Marketing
• Deep Learning
• Microsoft Office
• Microsoft Sql Server
• Brand Management
• Spss Analysis
• Natural Language Processing
• Knitting
• Fiber Testing
• Supply Chain Management
• Transact-Sql (T-Sql)
• Statistical Data Analysis
• Regression Analysis
• Photoshop
• Python
• Omni-Channel Experience
• Oil Painting
• Customer Relationship Management (Crm)
• Ecommerce
• Business Intelligence
• Customer Analytics
• Technical Textiles
• Decision Making
• Machine Learning
• Microsoft Products
• Weaving
• Apache Spark
• Data Mining
• Spinning