Building robust, production-ready machine learning pipelines is usually a far removed experience from participating in the highly sanitized world of Kaggle competitions and million dollar Netflix prizes. Input data often needs to be prepared before model training can begin, sophisticated algorithms that work on sample data sets need to be scaled to large production data sets, platforms & pipelines are needed to regularly and reliably retrain and redeploy models, a wide-range of SLAs from hours to seconds need to be met, etc… Aside from the practical challenges of productizing AI, what are latest innovations in the field? Come to this track to learn how leaders in the industry build innovative ML-driven applications & systems as well as learn about some of the latest advances in the field.
Track: The Practice & Frontiers of AI
Location: Seacliff ABC
Day of week:
Track Host: Shubha Nabar
ML for Question and Answer Understanding @Quora
Quora's mission is to share and grow the world’s knowledge. On Quora, people ask questions on a wide range of topics and Quora surfaces those questions to people with relevant credentials and experiences so they can respond with an insightful, helpful answer. The more you use Quora—whether it’s to ask a question, answer one, or follow people or topics of interest—the better Quora gets. We’re constantly improving our ability to personalize an experience that’s filled with people, questions, and answers you’ve shown interest in. We achieve this via several machine learning and NLP systems.
In this talk, I will discuss these machine learning and NLP systems in depth. I will explore how we extract intelligence from questions on Quora, including how we do question-topic labeling, how we automatically correct questions with bad spelling and grammar, how we detect duplicate questions, how we learn to rank answers to questions and more. I will explain how the output of these systems supply an important input for the downstream machine learning applications that power Quora. Finally, I will highlight lessons I have learned from applying state-of-the-art machine learning techniques to consumer products at scale.
Automating Netflix ML Pipelines With Meson
In this talk we discuss the evolution of ML automation at Netflix and how that lead us to build Meson, an orchestration system used for many of the personalization/recommendation algorithms. We will talk about challenges we faced, and what we learned automating thousands of ML pipelines with Meson.
Models in Minutes not Months: AI as Microservices
Companies are redefining their businesses by building models and learning from data. Whether it is using data science to predict their best sales and marketing targets, automating digital customer interactions using bots, or reducing waste in logistics and manufacturing - Artificial Intelligence will improve your business once deployed.
Serving up good predictions at the right time to drive the appropriate action is hard. It requires setting up data streams, transforming data, building models and delivering predictions. Most approach this by building single models and realizing along the way that data science is only the beginning. The engineering and infrastructure required to maintain a single model and ship the predictions present even more challenges.
Trying to replicate this success for more models or customers is even more difficult. Most approach it by building a handful of additional models, painstakingly addressing challenges by taking one-off approaches to handling increasing volumes of data, differences in data, changes in process, etc. Scaling to 1000s of customers becomes impossible.
At Salesforce we built the Einstein Platform to enable the automation and scaling of Artificial Intelligence to 1000s of customers, each with multiple models. The data ingestion, automated machine learning, instrumentation and intelligent monitoring and alerting make it possible to serve the varied needs of many different businesses. In this talk we will cover the nuts and bolts of the system, and share how we learned to solve for scale and variability with a fully operational Machine Learning platform.
Michelangelo: Uber’s Machine Learning Platform
Michelangelo is the Machine Learning platform that we have built at Uber. The purpose of Michelangelo is to enable data scientists and engineers (and eventually non-technical users) to easily build, deploy, and operate machine learning solutions at scale. It is designed to be ML-as-a-service, covering the end-to-end machine learning workflow: manage data, train models, evaluate models, deploy models, make predictions, and monitor predictions. Michelangelo supports traditional ML models, time series forecasting, and deep learning. In this talk, I will use one of our models, the UberEATS estimated delivery time model, as a case study to illustrate how the system works end-to-end. I will also cover some of the lessons we learned while developing and scaling the platform.
The Practice & Frontiers of AI Panel
Join the track speakers and invited guests as they discuss where AI is heading and how it's affecting software today.
Chris Moody, Manager of the Applied AI team @StitchFix
Reena Philip, Engineering Manager @Facebook
Kevin Moore, Senior Data Scientist @ Salesforce Einstein
Miju Han, Director of Product @GitHub
Melanie Warrick, Senior Developer Advocate for ML and Google Cloud
Building Bots and Conversational AI
People are spending a lot of time on messaging and voice conversational mediums such as Facebook Messenger and Amazon Alexa, which have opened up for building bots. These bots are allowing services and businesses to connect with users on these conversational interfaces. Conversational bots requires natural language processing, extracting relevant information, understanding context, and coming up with responses to users messages. Recent advances in deep learning has led to tremendous progress in natural language processing and is making conversational AI a reality. This talk will describe how to build a conversational bot, use deep learning for natural language processing, and deploy it on a conversational platform.
Last Year's Tracks
Monday, 1 November
-
Microservices / Serverless Patterns & Practices
Evolving, observing, persisting, and building modern microservices
-
Practices of DevOps & Lean Thinking
Practical approaches using DevOps & Lean Thinking
-
JavaScript & Web Tech
Beyond JavaScript in the Browser. Exploring WebAssembly, Electron, & Modern Frameworks
-
Modern CS in the Real World
Thoughts pushing software forward, including consensus, CRDT's, formal methods, & probabilistic programming
-
Modern Operating Systems
Applied, practical, & real-world deep-dive into industry adoption of OS, containers and virtualization, including Linux on Windows, LinuxKit, and Unikernels
-
Optimizing You: Human Skills for Individuals
Better teams start with a better self. Learn practical skills for IC
-
Open Spaces
Tuesday, 2 November
-
Architectures You've Always Wondered About
Next-gen architectures from the most admired companies in software, such as Netflix, Google, Facebook, Twitter, & more
-
21st Century Languages
Lessons learned from languages like Rust, Go-lang, Swift, Kotlin, and more.
-
Emerging Trends in Data Engineering
Showcasing DataEng tech and highlighting the strengths of each in real-world applications.
-
Bare Knuckle Performance
Killing latency and getting the most out of your hardware
-
Socially Conscious Software
Building socially responsible software that protects users privacy & safety
-
Delivering on the Promise of Containers
Runtime containers, libraries, and services that power microservices
-
Open Spaces
Wednesday, 3 November
-
Applied AI & Machine Learning
Applied machine learning lessons for SWEs, including tech around TensorFlow, TPUs, Keras, PyTorch, & more
-
Production Readiness: Building Resilient Systems
More than just building software, building deployable production ready software
-
Developer Experience: Level up your Engineering Effectiveness
Improving the end to end developer experience - design, dev, test, deploy, operate/understand.
-
Security: Lessons Attacking & Defending
Security from the defender's AND the attacker's point of view
-
Future of Human Computer Interaction
IoT, voice, mobile: Interfaces pushing the boundary of what we consider to be the interface
-
Enterprise Languages
Workhorse languages found in modern enterprises. Expect Java, .NET, & Node in this track