Ho Huang

About Ho Huang

Accomplished Data Scientist with 6 years of experience architecting and deploying production-grade machine learning systems, including recommendation, classification, and Gen AI-powered solutions. Proven track record of owning end-to-end ML pipelines, leading large-scale A/B experiments, and delivering measurable business impact in high-ambiguity environments. Expert in Python, SQL, PySpark, and AWS.Skillsets- Programming Skills: Python, SQL, Spark, Scala, R- Generative AI: LLMs (Claude, GPT), LangChain, Prompt Engineering, Knowledge Base, AWS Bedrock- ML & Stats: Regression, Classification, NLP, Deep Learning, Causal Inference, A/B Testing, Propensity Modeling- Cloud & Tools: AWS (SageMaker, Lambda), MLOps, Git, Docker, Tableau

Experience

1. Data Scientist Ii

   Amazon

   Jun 2024 — Present · Bellevue, WA, US

   Architected and productionized 3 end-to-end ETL and ML pipelines using Python, PySpark, AutoML, and AWS, enabling scalable model training, deployment, and monitoring across large-scale customer datasets• Developed propensity and segmentation models supporting online marketing campaigns, generating $9M+ in revenue and 170K units sold through improved customer targeting• Designed and deployed 5+ recommendation models using advanced similarity scoring and custom ranking logic, achieving an average 3% incremental lift in units sold across multiple customer-facing surfaces• Led a GenAI initiative using LLMs and LangChain to automate product attribute generation, eliminating hours of manual effort previously required for human curation• Revamped customer-facing classification models by introducing new data sources, modernized ETL pipelines, a BERT-based model and LLM-based reasoning, improving labeling accuracy and model explainability• Co-designed and analyzed 100+ A/B tests spanning CX changes, automation, and ML-driven solutions, in close partnership with cross-functional product and engineering leads• Launched a company-wide experiment metric to accurately measure bundle performance across experiments and standardize experiment evaluation• Conducted deep-dive cost–benefit analyses on social media advertising spend, identifying $17M in anomalous spend and influencing leadership-level decision making• Built causal inference and panel data models to evaluate homepage CX changes in scenarios where randomized experiments were not feasible, enabling reliable offline impact estimation• Mentored junior data scientists on ETL design, ML pipeline development, and best practices for production ML systems

Education

• National Taiwan University

  Bachelor's degree, Agriculture Economics

  2013 — 2018

• University of California, Berkeley

  Summer Session

  2015 — 2015

• Draper University

  Entrepreneurship Program

  2017 — 2017

• University of Maryland

  Master of Science - MS, Business Analytics

  2018 — 2019