This talk breaks down stage-by-stage requirements and challenges for online prediction and fully automated, on-demand continual learning. We’ll also discuss key design decisions a company might face when building or adopting a machine learning …
Managing the Flywheel of ML Data
The ML Engineer’s life has become significantly easier over the past few years, but ML projects are still too tedious and complex. Feature stores have recently emerged as an important product category within the MLOps ecosystem. They solve part of …
[Open Source] Hamilton, a micro framework for creating dataframes, and its application at Stitch Fix
At Stitch Fix, we have 130+ “Full Stack Data Scientists” who, in addition to doing data science work, are also expected to engineer and own data pipelines for their production models. One data science team, the Forecasting, Estimation, and Demand …
Twitter’s Feature Store Journey
A Feature Store is an essential piece of a production ML system. Twitter’s journey of building Feature Stores began several years ago. Since then, we have gone through multiple iterations of our Feature Store to facilitate creating, organizing, …
Managing Data Infrastructure with Feast
Feast provides a simple framework for defining and serving machine learning features. In order to serve features reliably, with low latency and at high scale, Feast relies heavily on cloud infrastructure such as DynamoDB or AWS Lambda. This talk …
Using Feast in a Ranking System
This will be a practical session explaining how Better.com uses Feast in a Ranking System that depends on multiple data sources and several models. We’ll provide a walkthrough of several architectures we considered as a team to manage features, …
Data Transfer Challenges In Evaluating MLOps Platforms
Customers evaluating MLOps platforms as a service need to provide customer data during the evaluation phase. The data often needs to be moved to the MLOps companies’ warehouses. This is not a simple task and can become costly if the two partners are …
A Tour of Features in the Wild and a Modern Solution to Manage Them
Mike will kick off the event and present his views on the different types of features commonly used for Operational ML use cases, and solutions to manage them. Operational ML models rely on several types of features with different characteristics …
What Data Engineers Should Know About Real-Time Analytics
Online Personalization is powered by a real-time feature engineering platform that ingests, processes, and joins millions of events per minute. We present the architecture that manages the real-time feature engineering platform at Personalization. …
Streaming Architecture with Kafka, Materialize, dbt, and Tecton
Drizly is building out our Data Science stack and streaming infrastructure to match the success we’ve had with the modern data stack on BI. We are currently standing up our architecture using Kafka, Materialize, and dbt. We are planning on adding …
ML Design Patterns for Data Engineers
As machine learning moves from being a research discipline to a software one, it is useful to catalog tried-and-proven methods to help engineers tackle frequently occurring problems that crop up during the ML process. In this talk, I will cover three …
