Operationalized ML is the coveted final step for data science organizations. Yet, most O16N success stories are still being told by companies born out of data. In this fireside chat, Dataiku will speak to leaders at Palo Alto Networks, Clorox, and Johnson & Johnson about their experiences and struggles in moving AI out of the sandbox and into production.
Algorithmia is MLOps software that manages all stages of the ML lifecycle. In this session, we'll demonstrate Algorithmia in action and show you how it solves common management challenges to deploy models, connect to data sources, automatically scale model inference, and manage the ML lifecycle in a centralized model catalog.
The landscape of ML tooling is becoming richer and richer, but the downside is a jungle of opinionated tooling in the ecosystem that can be overwhelming. Here we will take a look at how the tech giants are solving these problems, explore buy vs. build options and look at a template you need to consider for your own ML workflow.
Theator is a Surgical Intelligence Platform providing revolutionary personalized analytics on lengthy surgical operation videos. In this session, we will make a brief - but deep - dive into Theator's continuous training and inference pipelines, and detail the exciting interplay between in-house and provided.
Prosus is a global consumer internet group serving >1,5B people through its investments in food delivery, online classifieds, payments and edtech in growth markets like India, Russia and Brazil. This talk explores the evolution of ML Platform capabilities at several Prosus companies to enable applying machine learning at scale.
Traditional approaches to managing technical projects and teams or establishing organizational culture can be at odds with quick innovation and success, particularly with machine learning and AI. Here, we discuss how ML/AI executives can build highly effective ML teams, support them, and shift the broader organizational culture toward innovation and adoption of ML.
ML projects are expensive. There is a natural tension between business leaders investing in front-line ML projects, and the platforms, tools, and teams that can accelerate these projects. In this session, we discuss how to explain the value of ML platforms and infrastructure investments and how to build the business case for them.
As companies adopt AI/ML, they run into questions of forecasting cost and ROI. In this session, we will give a brief overview of Intuit's ML platform, with a specific focus on operational cost transparency across feature engineering, model training, and hosting. We will also talk about the benefits we've observed.
In her talk, Jennifer Prendki will dig into how the lack of attention from experts on data collection and preparation is the likely cause for a still highly dysfunctional ML lifecycle. She will explain how “Data Prep Ops” can change the game, and how to better incorporate concepts such as active learning, human-in-the-loop ML, and strategic data collection as integral parts of the ML lifecycle.
Fraud detection is frequently one of the most demanding real-time models, requiring real-time decisions made on real-time data inputs. In this workshop, we will demonstrate how you can use a feature store to build fraud detection features, build a training dataset to train a model, and then deploy that model with AWS SageMaker.
In this workshop, we explore how Machine Learning on the Cloudera Data Platform enables fast and effective end-to-end machine learning workflows -from the data prep and pipelines that power the models to managing the move to production, and finally how teams can achieve explainable and trustworthy predictive applications.
One of the biggest challenges executives face is ensuring project success. Far too often precious resources are wasted on projects that fail to generate value, don't get off the ground, or are inadequately resourced. This keynote explores the reasons for these failures and best practices for a successful launch and deployment of ML & AI solutions.
This summit begins with a series of short conversation starters by accomplished ML/AI leaders, followed by a Q&A and open discussion among Executive Summit participants Ameen Kazerouni. chief analytics officer of Orangetheory Fitness, Paul van der Boor, senior director of data science at Prosus Group and Hussein Mehanna vice president and head of ML/AI at Cruise.
In this hands-on demonstration, you will learn to use Splice Machine’s platform to build a feature store, train your favorite ML models on data from the feature store, and deploy them as an intelligent table inside our database. This workshop will conclude with an ML modeling competition and the opportunity for some Splice Machine swag!
Smartphones have seen explosive growth in AI/Machine Learning usage in the past 2-3 years. In this session, we will take you through the latest trends, opportunities, challenges, and tradeoffs associated with leveraging on-device AI acceleration to enable cutting-edge AI use cases on smartphones & other consumer devices.
Mike will provide an overview of feature stores - data platforms that make it easy to build, deploy, and use features in production- the key differences between types, and how to pick the right one based on your requirements. Then, join us as he shows how they helped Uber scale to 1000s of production models in just a few years.
Faisal is the Director of Engineering for Personalization Infrastructure at Netflix, running multiple teams delivering large-scale ML infrastructure. His team currently supports model development, model tools and data, and model serving. Here, Faisal and Sam will discuss lessons learned building Netflix's large-scale personalization infrastructure.
Ya Xu leads the LinkedIn data science practice, a team that touches every aspect of the organization from business investments to economic insights for policymakers, and much more. Here, Sam and Ya explore her career transition, the horizontal tools her team maintains, and some of the areas her team is leading innovation in.
There are 2 levels of real-time machine learning: 1) making online predictions in real-time, and 2) incorporating new data to update your model in real-time. This talk covers the state of real-time machine learning in production, as well as the staggering differences in its adoption across Internet companies in the US and China.
When many businesses start their journey into ML and AI, it's common to place a lot of focus on coding and data science algorithms. In reality, data science and models are only one part of the enterprise machine learning puzzle. In this presentation, you will learn how to effectively tackle the whole of that puzzle.
We typically hear conference presentations from the single perspective of an organization's data scientists, data engineers, platform engineers, or ML/AI leaders. "Team Teardown" turns this model on its head, speaking with several members of an organization's team. Today, we're diving into Driving Platform Adoption and Success at Spotify
We typically hear conference presentations from the single perspective of an organization's data scientists, data engineers, platform engineers, or ML/AI leaders. "Team Teardown" turns this model on its head, speaking with several members of an organization's team. This week, we'll be looking at MLOps in the Cloud for IoT at iRobot
If you’ve brought two or more ML models into production, you know the struggle that comes from the complex process. This talk will teach you a whole new approach to MLOps that allows you to successfully scale your models without increasing latency, by merging a database, a feature store, and machine learning.
Delivering a bad model into production/serving is deceptively easy to do. Using a hand analysis of approx 100 incidents, we identify common causes and manifestations of failures, provide some idea for how to measure the potential damage, and propose a set of techniques for detecting problems before they cause damage.