Skip to navigation Skip to main content

Registration & Breakfast

Main

Registration & Breakfast

Chair's opening remarks

Hugo Bowne-Anderson Data Scientist DataCamp

Opening Keynote: The future of machine learning

Every year, machine learning becomes more sophisticated and we come closer to unlocking its true potential. Sarah Hoffman draws on her wealth of experience to discuss the future of machine learning and how organizations will need to transform to work with this technology.

Sarah Hoffman VP, AI and Machine Learning Research Fidelity Investment

Sparking more Meetups with machine learning

Meetup brings people together in real life to share their interests and passions. Matching newly created Meetup groups with interested members is a key part of that platform. In this talk, I will share the practices we followed in modernizing Meetup’s approach to that recommendations problem. You’ll learn how we build and operate a robust and scalable data and recommendations system that notifies millions of members each day, how we measure and monitor that system, and how we apply an iterative product development cycle to best improve user experience driven by machine learning.

Zachary Cohn Principal Machine Learning Engineer Meetup

Customized regression model for Airbnb dynamic pricing

We describe the pricing strategy model deployed at Airbnb, an online marketplace for sharing home and experience. The goal of price optimization is to help hosts who share their homes on Airbnb set the optimal price for their listings. In contrast to conventional pricing problems, where pricing strategies are applied to a large quantity of identical products, there are no “identical” products on Airbnb, because each listing on our platform offers unique values and experiences to our guests. The unique nature of Airbnb listings makes it very difficult to estimate an accurate demand curve that’s required to apply conventional revenue maximization pricing strategies. In this talk, I will give an overview of the pricing system we use at Airbnb.

Hangjun Xu Senior Machine Learning Software Engineer Airbnb

Networking coffee break

BREAKOUT SESSION - splits into 2 rooms

The room will split into two rooms, choose one topic to participate in:

1. Machine learning in the enterprise: creating your game plan, sponsored by Oracle /
2. Pragmatic AI: From aspiration to actionable strategy to deployed performance driving solutions, sponsored by SAP

Andreas Welsch Regional Business Development Lead SAP.
Chuck Hollis SVP Oracle

Panel: How do we build more 'AI savvy' teams?

Zachary Cohn Principal Machine Learning Engineer Meetup
Sanksshep Mahendra Head of Technology Merriam-Webster
William Tran Machine Learning Engineer Pager

Networking lunch

Machine learning for product discovery in E-Commerce

Artificial Intelligence is revolutionizing all aspects of e-commerce. In this talk we will focus on product discovery, which encompasses the data and the tools empowering our customer to quickly and effortlessly find exactly what they are looking for. We will discuss how Macys and Bloomingdales brands use Machine Learning techniques from deep learning image models to sophisticated collaborative shopping algorithms in order to shorten the customer’s journey from a shopping intent to a purchase. We will highlight the role Macys Enterprise Machine Learning Platform plays in reducing the time to market for hottest research projects delivered by Macys Lab

Denis Kamotsky Principal Engineer for Search and Machine Learning Macy's

Machine learning in the enterprise: A capabilities overview

The full impact of machine learning will eventually impact every role and function in the enterprise. Fully exploiting machine learning will require addressing the needs of many groups in the enterprise, from business users to data scientists.

Join Chuck Hollis as he maps out Oracle’s approach to capitalizing on machine learning by targeting the distinct needs of six different enterprise personas.

Chuck Hollis SVP Oracle

Personalized recommendations for food, recipes, and groceries

Plated face a unique problem of assigning weekly recommendations of novel food recipes to both new and old customers. In this talk, Andrew Marchese will discuss how they derive features from extremely content-rich recipes, and how these features are used to drive personalized recommendations that learn customer affinity towards certain feature sets. Additionally, learn how this process is scaled in order to recommend the recipes to a (much) larger set of grocery store customers, and how it can be extended to recommend any grocery product.

Andrew Marchese Data Scientist Plated

Networking coffee break

Using machine learning for innovative methods of demand forecasting

Inventory costs for service parts tie up in excess of two billion dollar per year in working capital only. It is pervasive that service parts are forecasted based on past orders data. This is essentially passively waiting for the history to repeat itself. During periods of instability and high volatility, this can lead to delayed response in the inventory levels. In cases of increasing demand, this delay can cause loss of sales, delayed revenue realization, and loss of customer loyalty. In cases of decreasing demand, this can yield sub-obtimal inventory carrying costs. We can be more proactive while extrapolating the past into the future. Machine telematic prognostics when aggregated appropriately can serve as a strong leading indicator to logistics planning and parts demand forecasting. Machine failure prognostics and network optimization can be leveraged to move and stock the parts closer to the probable failure.

Seifu Zerihun Analytics Team Leader Caterpillar

Applications of natural language understanding in patient-facing healthcare

Recent advances in deep learning have enabled a whole new range of Machine Learning models in the field of Natural Language Processing and Understanding. In this talk, William will talk about how Pager is applying these cutting-edge techniques to our care-coordination chat platform to surface relevant patient information based on the current chat context and assist our nurses in triaging patients with more confidence. This results in faster chat response, lower patient wait time, and increased nurse productivity.

William Tran Machine Learning Engineer Pager

Chair's closing remarks

Networking drinks reception

Registration and Breakfast

Main

Registration & Breakfast

Chair's opening remarks

Sanksshep Mahendra Head of Technology Merriam-Webster

How to build a lean recommendation engine

As businesses continue to explore and invest in machine learning, they will need to be smart and conscientious about how they collect and use data to enhance their products or drive a bottom line. In this talk, we will use a case study on personalizing people's news feeds to introduce an agile framework for building, evaluating, and iterating algorithmic solutions.

Harsha Rao Senior Product Manager WeWork

Panel: Democratization of AI - making data accessible to everyone

Allison Gorbaty Senior Director of Marketing Crunch
Carl Anderson Director, Data Science WW (formerly Weight Watchers)
William Tran Machine Learning Engineer Pager
Sanksshep Mahendra Head of Technology Merriam-Webster

Machine Learning on Distributed Systems

Most real-world data science workflows require more than multiple cores on a single server to meet scale and speed demands, but there is a general lack of understanding when it comes to what machine learning on distributed systems looks like in practice. Gartner and Forrester do not consider distributed execution when they score advanced analytics software solutions. Many formal machine learning training occurs on single node machines with non-distributed algorithms. In this talk we discuss why an understanding of distributed architectures is important for anyone in the analytical sciences. We will cover the current distributed machine learning ecosystem. We will review common pitfalls when performing machine learning at scale. We will discuss architectural considerations for a machine learning program such as the role of storage and compute.

Waqas Dhillon Product Manager - Machine Learning Microfocus

Networking break

How AI is transforming cancer

Much has been said about the potential impact of Artificial Intelligence in the healthcare domain in general and in the cancer space in particular.
Yet, real case studies that show the impact of AI on patients as well as on the business are lacking.
This presentation focuses on a case study at MSK where integrating AI into the clinical workflow saved not only lives but also reduced the cost of care.

Isaac Wagner Director of Strategy Analytics Memorial Sloan-Kettering Cancer Center

Equity, Inequity, & Machine Learning

Carta is mapping one of the world's most valuable propriety datasets: the ownership graph. In this talk, I'll present recent work analyzing how startups distribute ownership across their employees, and discuss the implications for companies looking to deploy more just and equitable compensation strategies. Then, I'll sketch some fundamental questions machine learning can answer about ownership, and outline strategies for tackling them at scale.

Jerry Talton Director of Data and Machine Learning Carta

Roundtable discussion

Networking lunch

Developing a food recognition software for targeted weight loss

What does it take to develop an algorithm that accurately recognizes food through your phone? Focusing on the user experience, Lose It! was able to build software to suggest foods, as well as show the calorie count and nutritional information, all through a photo. Will Lowe will share how he developed the revolutionary photo recognition software to maximize their consumer experience.

Will Lowe Senior Data Scientist Lose It!

Making the machines work for us

At Dia&Co, a plus-size women's styling service, we have a classic human-in-the-loop process wherein personal stylists interact with both machine learning models and our customers. Such a process faces a tradeoff between automation and autonomy. This talk will detail how we have optimized for letting both woman and machine do what they do best. Through the lens of a case study around a recent product launch, I'll cover the choices we made in designing the loop for this new product. This will include both the machine learning and operations research models used and how we fit these models into the loop. I'll close with a discussion of how this loop is designed to improve over time in a virtuous cycle.

Ethan Rosenthal Lead Machine Learning Engineer Dia & Co.