• Registration & Breakfast

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 7:30 am 8:30 am

  • Introduction to Deep Learning (8:30AM-5:00PM)

    LOVELACE STAGE- LEVEL 2 : 8:30 am 5:00 pm

    Lukas Biewald, Founder, Figure Eight |

    This a course to take engineers from zero to one in machine learning in a day. The world needs more people that understand machine learning and our goal is to get you started on that path as efficiently as possible. Learn more

  • Executive Briefing AM Session (10:00AM-12:00PM)

    LOUNGE - LEVEL 2: 10:00 am 12:00 pm

    Robin Bordoli, CEO, Figure Eight |

  • Lunch

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 12:00 pm 1:00 pm

  • Executive Briefing PM Session (3:00PM-5:00PM)

    LOUNGE - LEVEL 2: 3:00 pm 5:00 pm

    Robin Bordoli, CEO, Figure Eight |

  • Opening Reception (5:00PM-6:30PM)

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 5:00 pm 6:30 pm

  • Registration & Breakfast

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 7:30 am 9:00 am


    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 10:30 am 10:45 am

  • 3 Sessions

    A Gif is Worth a 1000 Words: Animated Thoughts About the Future of Enterprise Intelligence

    AL-KHWARIZMI STAGE - LEVEL 2: 10:45 am 11:05 am

    Anthony Johnson, CTO, GIPHY |

    This talk covers the many deep learning/machine learning and big data problems underlying searching for video at scale, and why despite the fact that GIFs are the furthest thing you can imagine from the enterprise, what we do today is probably how you are going to access data in the enterprise in a few years.

    Challenges & Design Patterns for Conversational AI

    TURING STAGE - LEVEL 2: 10:45 am 11:05 am

    Peter Skomoroch, Co-Founder and CEO, SkipFlag (acquired by Workday) |

    Enterprise and consumer applications increasingly apply machine learning to create conversational interfaces. Adding a conversational UX presents a number of challenges for machine learning practicioners attempting to build intelligent applications. This session will describe some lessons learned from the recent wave of bots and from building SkipFlag, an intelligent knowledge base that integrated with Slack. Should you develop your own algorithms or make use of NLP as a service? How should you plan to include humans in the loop? To what degree do you need to specialize your models for the industry you’re working in? We'll cover these questions and more.

    Customer-Focused Machine Learning for the Creative Landscape

    LOVELACE STAGE - LEVEL 2: 10:45 am 11:05 am

    Alex Filipkowski, Product Manager, Applied Science and Machine Learning, Adobe |

    At Adobe, we’ve started a new applied science team tasked with using research-inspired methods to solve real-world user problems. We train our models with native content and custom annotations, which enables us to build deep learned machinery tailored specifically to the use case at hand.

  • 3 Sessions

    Beyond ImageNet: Building "AI Ready" Datasets in Noisy, Real-World Environments

    AL-KHWARIZMI STAGE - LEVEL 2: 11:15 am 11:35 am

    Mikel Rodriguez, Director of Computer Vision Research , MITRE |

    Fueled by large-scale labeled datasets like ImageNet, in the past few years computer vision has evolved from the realm of science fiction to a vast array of applications that include everything from healthcare to transportation. However, getting labeled training data has become the key development bottleneck in supervised machine learning. As we move beyond ImageNet, the challenge of building "AI ready" datasets is even greater in the noisy, unstructured and highly biased data distributions that characterize most real-world applications. This talk will present several case studies exploring how we can jumpstart the machine learning process.

    Is Science Just a Game?

    TURING STAGE - LEVEL 2: 11:15 am 11:35 am

    Adrien Treuille, VP of Simulation, Zoox |

    This talk describes Foldit and EteRNA, a series of scientific discovery games we have developed to solve real scientific problems. These games lead us to wonder: how many unknown "Kasparovs" are out there on the Internet -- potential experts at tasks they never knew existed? Is this the future of expertise?

    Notes from the Field: The Platform, People and Processes of Agile Data Science

    LOVELACE STAGE - LEVEL 2: 11:15 am 11:35 am

    Sarah Aerni, Director of Data Science , Salesforce |

    As data scientists, our business users ask us to produce the perfect model to supercharge their products. While platforms and processes evolved to address challenges in software development, for data science this is still very nascent. In this talk I will share how Salesforce tackles the challenge of making data science an agile partner to over 100,000 customers without compromising on high quality AI solutions. I will cover both the process, and the platform that enables experimentation and deployment to enable each of our customers to build custom models with a few simple clicks.

  • 3 Sessions

    Modeling AI Advances Through Data

    LOVELACE STAGE - LEVEL 2: 11:40 am 12:00 pm

    Jack Clark, Strategy & Communications Director, OpenAI |

    Using real data (and lots of it), this session measures the real progress of AI through the years showcasing AI adoption, the AI Index and multiple additional data sets that truly benchmark real world usage of AI

    Productizing Deep Learning for the Enterprise

    TURING STAGE - LEVEL 2: 11:40 am 12:00 pm

    Hanlin Fang, Director of Machine Learning
    Product Management, Workday |

    When consumers experience AI/ML benefit from various sources in our daily life, enterprises are facing challenges when applying similar AI/ML techniques to transform business. In this session, we will share how Workday (Enterprise SaaS company on HCM and FIN) has identified specific business problem for ML to solve, collected enough data to prototype, and deployed the solution as part of Workday Application product available to all Workday customers in less than 18months. We will also share lesson learned from legal, privacy, and security aspect with Human-in-the-loop approach which is a critical part of enterprise ML product development journey.

    Reality Check: Beyond the Hype. Real Companies Doing Real Business Getting Real Value with AI

    AL-KHWARIZMI STAGE - LEVEL 2: 11:40 am 12:00 pm

    Alyssa Simpson Rochwerger, VP of Product , Figure Eight |

    AI is like sex in high school - everyone is talking about it but who is actually doing it (and generating business results). This session takes an industry by industry perspective on true AI adoption disambiguating the hype from the reality, the theoretical from the practical and the research labs from ROI.


    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 12:00 pm 1:00 pm

  • Figure Eight Customer Showcase

    TURING STAGE - LEVEL 2: 1:00 pm 2:00 pm

    Daniel Golden, Director of Machine Learning, Arterys | Etienne Manderscheid, Head of Data Science and Co-Founder, TalkIQ | Diego Represas, Data Science, | Estelle Afshar, Senior Manager, Data Science Online, Home Depot | Kevin Hu, Data Scientist, Plaid | Chris Padwick, Deep Learning Team Lead, Blue River Technology |

    Always a highlight of Train AI, the customer showcase is a rapid-fire session featuring some of the most innovative use-cases of AI. Presenting companies include: Dan Golden- Director of Machine Learning, Arterys; Kevin Hu- Data Analyst, Plaid; Etienne Manderscheid- Head of Data Science and Co-founder, Talk IQ; Diego Represas- Data Science, Digit; Estelle Afshar- Senior Manager Data Science Online, Home Depot

  • 2 Sessions

    Learning for the Long Tail

    TURING STAGE - LEVEL 2: 2:05 pm 2:25 pm

    Dr. Marti Hearst, Professor, UC Berkeley |

    We are building a next-generation system to help readers better understand today's chaotic news landscape. This project combines information integration, natural language processing (NLP), and information visualization in a novel way, thus testing the limits of existing NLP algorithms in a real world context. In this talk we will illustrate an advance on a long-tail problem, that of entity linking. Our system is skilled at determining which wikidata a person refers to, but also, more unusually, it can tell when a person mention is likely *not* to refer to someone in the knowledge base. This talk will include a live demo of the system.

    The Unreasonable Effectiveness of Training Data

    LOVELACE STAGE - LEVEL 2: 2:05 pm 2:25 pm

    Pete Warden, Engineer, Google |

    This talk will cover why good training data is so important in machine learning, what makes data ""good"", and how to improve your process.

  • 2 Sessions

    AI Ethics Panel Discussion: Beyond Killer Robots and Trolley Problems

    TURING STAGE - LEVEL 2: 2:30 pm 3:10 pm

    Blair Hanley Frank, Principal Analyst, ISG | Ruchir Puri, IBM Fellow and Chief Architect, IBM Watson | Lisha Li, Principal, Amplify Partners | Radha Basu, CEO, iMerit Technology Services | Tarin Ziyaee, Founder, "Stealth" |

    The long-term existential threats of AI have taken over our popular consciousness, but what about the everyday issues practitioners and the public face right now? This panel will explore the current ethical challenges in the AI field, with an eye towards how they might impact the future.

    From PhD to Product: Panel Discussion

    LOVELACE STAGE - LEVEL 2: 2:30 pm 3:10 pm

    Robert Munro, Chief Technology Officer, Figure Eight | Anima Anandkumar, Principal Scientist at Amazon AI and Bren Professor at Caltech, Amazon Web Services | Kapil Gupta, Data Science Lead, Airbnb | Olya Gurevich, Independent Natural Language Scientist , |

    Leaders in AI applications will talk about their personal paths from being research-focused grad students to results-focused product leaders. They will share lessons learned from which parts of academia did (and did not) carry over to making AI work in the real-world, and provide guidance to people pursuing a similar path.

  • PM Break

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1: 3:15 pm 3:30 pm

  • TRAIN AI Closing Party (5:00PM-7:00PM)

    SPONSOR EXPO / TRAIN AI LOUNGE - LEVEL 1 : 5:00 pm 7:00 pm