cv

General Information

Full Name Edwin Ng
Email edwinnglabs@gmail.com
LinkedIn edwinng-ds
GitHub edwinnglabs

Experience

  • 2026 - Present
    Staff Data Scientist
    Instacart (Maplebear, Inc.), San Francisco, CA
    • Lead marketing planning with marketing, finance, and senior leadership to define vision, roadmap, and strategy for marketing science initiatives.
    • Build an in-house marketing measurement system unifying MMM, multi-touch attribution, and incrementality testing for reporting and budget allocation.
    • Develop LTV and MTA models to inform customer acquisition and channel investment decisions.
  • 2023 - Present
    Founder
    EdanSoul Analytics, Remote
    • General consulting in data science and strategy, focused on marketing and forecasting science.
  • 2022 - 2026
    Senior Applied Scientist
    Amazon.com Inc, San Francisco, CA
    • Led end-to-end development of a global forecasting system for Amazon Grocery outbound forecasting, driving inventory planning and replenishment across online and physical stores.
    • Led development of an explainable forecast system for Amazon inbound inventory, driving labor planning and purchasing decisions.
    • Led geo-level hierarchical marketing mix modeling across Amazon Grocery and Whole Foods Market.
    • Hosted talks and workshops, and mentored junior scientists within the forecasting science team.
  • 2020 - 2022
    Manager II, Applied Science
    Uber Technologies Inc, San Francisco, CA
    • Managed a team of applied scientists building scalable data-driven solutions for marketing measurement and growth strategy across drivers, riders, eaters, and couriers.
    • Led a customer value forecasting platform underpinning return-on-advertisement-spend (ROAS) decisions.
    • Led a centralized market-level test platform for experimentation design and causal inference using synthetic controls and interrupted time series.
  • 2019 - 2020
    Senior Applied Scientist I / TLM
    Uber Technologies Inc, San Francisco, CA
  • 2017 - 2018
    Applied Scientist I, II
    Uber Technologies Inc, San Francisco, CA
  • 2015 - 2016
    Data Scientist
    Ten-X.com (formerly Auction.com), Belmont, CA
    • Improved daily email open rates by 50% with a tree-based uplift model and look-alike algorithm.
    • Built and maintained pipelines and dashboards to monitor auction performance and model health.
  • 2014 - 2015
    Data Scientist
    PayPal Holdings, Inc., San Jose, CA
    • Developed a synthetic-control time-series model to quantify the impact of policy changes across markets.
    • Automated pipelines and dashboards tracking PayPal Digital Gift Card store performance.
  • 2013 - 2014
    Quantitative Developer
    Tanius Technology, LLC, Alamo, CA

Education

  • 2013
    M.S. in Statistics
    University of California, Los Angeles
  • 2010
    M.S. in Financial Engineering
    University of California, Los Angeles
  • 2009
    B.S. in Mathematics and Economics, Minor in Statistics
    University of California, Los Angeles

Open Source Projects

  • 2020 - Present
    Orbit: A Python package for object-oriented Bayesian time-series models
    GitHub
  • 2019 - Present
    Rlgt: Bayesian exponential smoothing models with trend modifications
    CRAN

Selected Talks

  • 2024
    Forecasting Grocery Demand with Bayesian State-Space Models and Causal Study
    KDD 2024, 3rd Customer Journey Workshop (Invited Speaker)
    Link
  • 2022
    Predicting Customer Lifetime Value Using Recurrent Neural Nets
    KDD 2022, 1st Customer Journey Workshop (Invited Speaker)
    Link
  • 2021
    Bayesian Time Varying Coefficient Model with Applications to Marketing Mix Modeling
    KDD 2021, AdKDD
    Link
  • 2021
    Orbit - An Object-Oriented Python Package for Time-Series Models
    The 41st International Symposium on Forecasting (ISF)
    Link
  • 2020
    Probabilistic Forecasting Applications at Uber
    The 40th International Symposium on Forecasting (ISF)
    Video
  • 2020
    Orbit: Probabilistic Forecast with Exponential Smoothing
    StanCon 2020
    Link

Publications

  • Chen, Ng, Smyl, Steininger. "Predicting Customer Lifetime Value Using Recurrent Neural Net" arXiv preprint arXiv:2412.2029 (2024)
  • Ng, Wang and Dai. "Bayesian Time Varying Coefficient Model with Applications to Marketing Mix Modeling" arXiv preprint arXiv:2106.03322 (2021)
  • Ng, Wang, Chen, Yang and Smyl. "Orbit: Probabilistic Forecast with Exponential Smoothing" arXiv preprint arXiv:2004.08492 (2020)

Honors and Achievements

  • Uber Eng Blog: Introducing Orbit, an open-source package for time-series inference and forecasting (2021)
  • Top 10% (Bronze Medal) with Team "CEZ" in the M5 Forecasting - Accuracy competition (2020)