Discover how Vital powers its predictive, customer-facing, emergency department wait-time product with request-time input signals and how it solves its “cold-start” problem by building machine-learning feedback loops using Tecton.
DIY Feature Store: A Minimalist’s Guide
A feature store can solve many problems, with various degrees of complexity. In this talk I’ll go over our process to keep it simple, and the solutions we came up with.
Workshop: Operationalizing ML Features on Snowflake with Tecton
Many organizations have standardized on Snowflake as their cloud data platform. Tecton integrates with Snowflake and enables data teams to process ML features and serve them in production quickly and reliably, without building custom data pipelines. …
More ethical machine learning using model cards at Wikimedia
First proposed by Mitchell et al. in 2018, model cards are a form of transparent reporting of machine learning models, their uses, and performance for public audiences. As part of a broader effort to strengthen our ethical approaches to machine …
Training Large-Scale Recommendation Models with TPUs
At Snap, we train a large number of deep learning models every day to continuously improve the ad recommendation quality to Snapchatters and provide more value to the advertisers. These ad ranking models have hundreds of millions of parameters and …
Intelligent Systems with real-time ML systems
In an omni-commerce space such as Walmart, Personalization is the key to enable customer journeys tailored to their individual needs, preferences and routines. Moreover, in e-commerce, customers’ needs and intent evolve with time as they navigate …
Machine Learning in Production: What I learned from monitoring 30+ models
It’s a software monitoring best practice to alert on symptoms, not on causes. “Customer Order Rate dropped to 0” is a great alert: it alerts directly on a bad outcome. For machine learning stacks, this means we should focus monitoring on the output …
Wild Wild Tests: Monitoring Recommender Systems in the Wild
As with most Machine Learning systems, recommender systems are typically evaluated through performance metrics computed over held-out data points. However, real-world behavior is undoubtedly nuanced, and case-specific tests must be employed to ensure …
Feature Engineering at Scale with Dagger and Feast
Dagger or Data Aggregator is an easy-to-use, configuration over code, cloud-native framework built on top of Apache Flink for stateful processing of real-time streaming data. With Dagger, you don’t need to write custom applications or manage …
Weaver: CashApp’s Real Time ML Ranking System
In this session, we will talk about one of the core infrastructure systems to personalize the experience on the CashApp, Weaver, and the work we did to scale it. Weaver is our real-time ML ranking system to rank items for search and recommendation …
Empowering Small Businesses with the Power of Tech, Data, and Machine Learning
Data and machine learning shape Faire’s marketplace – and as a company that serves small business owners, our primary goal is to increase sales for both brands and retailers using our platform. During this session, we’ll discuss the machine …
Streamlining NLP Model Creation and Inference
At Primer we deliver applications with cutting-edge NLP models to surface actionable information from vast stores of unstructured text. The size of these models and our applications’ latency requirements create an operational challenge of deploying …
