Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

29
lding Real Time Targeting Capabilit Capital One | Fast Marketing July 20, 2016 | H20 Open Tour | NYC

Transcript of Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Page 1: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Building Real Time Targeting Capabilities Capital One | Fast Marketing

July 20, 2016 | H20 Open Tour | NYC

Page 2: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Ryan ZottiSenior Data Engineer

Subbu ThiruppathySenior Software Engineer

EXPERTISE

FUN FACT

Big Data, Python, R, Java, Machine Learning, AWSFUN FACT

Big Data, Java, AKKA Play, AWS

Built a self driving remote controlled car

Recipient of Capital One’s most prestigious honor

EXPERTISE

Page 3: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One
Page 4: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One
Page 5: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

http://www.tibco.com/blog/2015/05/26/upcoming-webinar-integration-as-the-foundation-of-fast-data-may-28-2/

Page 6: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Source: http://thumbs.dreamstime.com/x/65-miles-per-hour-7772157.jpghttps://morganalyx.wordpress.com/2013/02/22/assertive-driving/

Page 7: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

http://www.dailymail.co.uk/health/article-2467478/What-causes-dry-eye-syndrome-cure-treatment.html

Page 8: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

MODEL DATA

MODEL DEPLOYMENT

MODEL SCORING

MODEL TRAINING

Our challenge is…

…striving to be fast at everything

Page 9: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Most current, up-to-date data

Available as soon as it’s ready

Low latency at scale

FAST MODEL DATA

Page 10: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Most current, up-to-date data

Available as soon as it’s ready

Low latency at scale

FAST MODEL DATA

Page 11: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Most current, up-to-date data

Available as soon as it’s ready

Low latency at scale

FAST MODEL DATA

Page 12: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Most current, up-to-date data

Available as soon as it’s ready

Low latency at scale

FAST MODEL DATA

Page 13: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Distributed computing to crunch data fast

Elastic scaling with the public cloud

Speed from parallelism

FAST MODEL TRAINING

Page 14: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Distributed computing to crunch data fast

Elastic scaling with the public cloud

Speed from parallelism

FAST MODEL TRAINING

Page 15: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Distributed computing to crunch data fast

Elastic scaling with the public cloud

Speed from parallelism

FAST MODEL TRAINING

Page 16: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Distributed computing to crunch data fast

Elastic scaling with the public cloud

Speed from parallelism

FAST MODEL TRAINING

Page 17: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Model adapts to evolving customer landscape

Automatically refit the model and daily deploy

Seamlessly integrate with existing Java tech stack

FAST MODEL DEPLOYMENT

Page 18: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Model adapts to evolving customer landscape

Automatically refit the model and daily deploy

Seamlessly integrate with existing Java tech stack

FAST MODEL DEPLOYMENT

Page 19: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Model adapts to evolving customer landscape

Automatically refit the model and daily deploy

Seamlessly integrate with existing Java tech stack

FAST MODEL DEPLOYMENT

Page 20: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Model adapts to evolving customer landscape

Automatically refit the model and daily deploy

Seamlessly integrate with existing Java tech stack

FAST MODEL DEPLOYMENT

Page 21: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Response < 100 milliseconds

JVM-based model (i.e. POJO)

Predictive power vs. runtime complexity (speed)

Gradient boosting provided the best balance

FAST MODEL SCORING

Page 22: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Response < 100 milliseconds

JVM-based model (i.e. POJO)

Predictive power vs. runtime complexity (speed)

Gradient boosting provided the best balance

FAST MODEL SCORING

Page 23: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Response < 100 milliseconds

JVM-based model (i.e. POJO)

Predictive power vs. runtime complexity (speed)

Gradient boosting provided the best balance

FAST MODEL SCORING

Page 24: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Response < 100 milliseconds

JVM-based model (i.e. POJO)

Predictive power vs. runtime complexity (speed)

Gradient boosting provided the best balance

FAST MODEL SCORING

Page 25: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Response < 100 milliseconds

JVM-based model (i.e. POJO)

Predictive power vs. runtime complexity (speed)

Gradient boosting provided the best balance

FAST MODEL SCORING

Page 26: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

VISITOR WEBSITE API MODEL DATA

Page 27: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Explore new technologies continuously

Ability to switch new models “on-the-fly”

Make the API faster

Incorporate new data sources

Resiliency, failover capabilities

Page 28: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

Technology changes

Flexibility of the cloud

Keep it simple

Small empowered teams

Page 29: Building Real Time Targeting Capabilities - Ryan Zotti, Subbu Thiruppathy - Capital One

THANK YOU