Sven Kreiss, Lead Data Scientist, Wildcard
summary
A production microservice takes a URL, returns JSON-structured article fields using a pipeline built on Apache Spark for batch training, PhantomJS scraping for rendered DOM features, and scikit-learn random forests for single-machine inference; labeling is partly human-powered. An iteration that replaces sampling with a feed‑forward neighbor-aware pass yields a 10x speedup with no quality loss and currently powers 54% of cards in the Wildcard iOS app.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:51What is Wildcard
- 6:54Features Pipeline
- 12:20Machine Learning
- 15:00Previous Approaches
- 16:35New Approach
- 17:25Conditional Random Fields
related talks
30mSmart Streams: A Real-time framework for scoring Big and Fast Data
01h 26mData Driven Action: a primer on Data Science
52mMachine Learning Data Pipelines
44mWildbook: Software and Machine Learning Engineers Join the Fight...
36mMachine Learning Microservices in Rails
38mFantastic ML apps and how to build them
01h 17mOne workshop that data scientists don't want you to attend...
01h 29mFull-stack Machine Learning and Generative AI for Data Scientists