Join us on Thursday, April 6th, from 12:00 pm to 6:30 pm, at our Global Headquarters in New York City, for an afternoon of thoughtful presentations and fast-paced roundtable discussions.


April 6, 2017

12 PM - 6 PM EST


AppNexus HQ

28 W. 23rd Street, New York

Join us in New York City for AppNexus Optimize

AppNexus Optimize provides a forum for engineers, data scientists and product specialists who want talk about ad tech with each other in all its depth and nuance – while avoiding the usual salesmanship and buzzwords. Whether you’ve attended Optimize before, or you’re joining us for the first time – don’t come expecting generalities.

Join us on Thursday, April 6th, from 12:00 pm to 6:30 pm, at our Global Headquarters in New York City, for an afternoon of thoughtful presentations and fast-paced roundtable discussions. Lunch and cocktails will be served.

We’re looking forward to seeing you there!


12:00 PM – 1:00 PM
Networking Lunch & Registration

1:00 PM – 1:20 PM
  • Brian O’Kelley
    CEO and Co-Founder
1:20 PM – 1:50 PM
Revolutionizing Click prediction with Appnexus programmable bidder v8
  • Brian O’Kelley
    CEO and Co-Founder
  • Abraham Greenstein
    Director of Data Science
  • Uri Bushey
    Technical Services Lead
1:50 PM – 2:20 PM
How Schibsted Media Group Balanced Scale and Real-Time Performance
  • Eugene Stipp
    VP, Engineering
2:20 PM – 2:50 PM
FULL STEAM A<HEAD>: two years in the header and still innovating
  • Matt Kendall
    Engineering Manager
  • Matt Jacobson
    Product Manager
2:50 PM – 3:10 PM
Discussion Topics Overview
  • Ben John
  • Ron Lissack
    Chief Architect
3:10 PM – 3:30 PM
Networking Break

3:35 PM – 4:20 PM
Discussion Topics Round 1

4:30 PM – 5:15 PM
Discussion Topics Round 2

5:15 PM – 6:30 PM
Cocktail Reception

Breakout Sessions

 The Client Insights Analytics (CIA) team partners with our biggest strategic clients to ensure that they are getting the most out of the data available to them, helping them reach their potential as advanced trading teams who confidently use log-level data to find interesting stories and actionable insights. This talk will provide an overview of the team’s philosophy, how CIA works with enterprise clients’ data science teams, and how we guide advanced trading in the programmable age.

Learn how to adopt GraphQL in your current architecture without rewriting your entire API library. We'll also dive into using GraphQL in client interfaces, and show you how a strongly-typed API can expedite your development and help you build faster applications. 

Find out how we’re leveraging YieldEx-related forecasting technology and other capabilities to enable AppNexus platform forecasting for buyers and sellers. We’ll spend time discussing the engineering challenges and technology involved, as well as the scope, direction, and future of forecasting technology.

Discover the dark and mysterious inner-workings of the virtual machine that powers the new logit custom models. We will provide attendees with an overview of its machine code, as well as details on its implementation. We will also compare it to the bonsai virtual machine and dive deep into how it’s able to accomplish more than regression models. 

When we started the video advertising initiative at AppNexus, we knew that our video player framework would need to power a variety of different products, and run on a multitude of platforms. Those imperatives made it clear from the get-go that we should build our player as modularly as possible. During this session, we will take a deep dive into our core video player architecture and discuss the lessons learned during the development process. While we will be discussing topics specific to our video player experience, these software engineering best practices can be applied to any type of project. 

The latest AppNexus optimization system is designed to focus on results — buyers can simply input their desired business outcomes, and our backend system will work to automatically achieve them. To accomplish this, we need a system of algorithms and machine learning models that work across a variety of use cases. In this talk, we'll walk through how our data scientists and engineers have broken the problem into manageable pieces and how the solutions fit together to make optimization “just work”. 

Join this session to learn how our internal Data Science and Optimization Engineering teams utilize the Genie Custom Model in Bidder to customize Click Prediction with log-level data and Logistic Regression models. We will walk through the general model training architecture, and review implementation suggestions that can help you achieve better prediction accuracy while still maintaining scalability. 

 During this session, we’ll be introducing new features like certificate-based authentication and Access Control. Additionally, we’ll describe what we have done to improve API stability over the last six months, and how we are keeping ourselves honest when it comes to monitoring.

Optimize Speakers

Uri Bushey

Technical Services Lead, AppNexus

Abraham Greenstein

Director, Data Science, AppNexus

Matt Jacobson

Product Manager, AppNexus

Ben John

CTO, AppNexus

Matt Kendall

Engineering Manager, AppNexus

Ron Lissack

SVP, Chief Architect, AppNexus

Brian O’Kelley

CEO and Co-Founder, AppNexus

Eugene Stipp

VP, Engineering, Schibsted Media Group