Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

Agile Data Science - Achieving Salesforce-Scale Machine Learning in Production

Association for Computing Machinery (ACM) via YouTube

Overview

Explore the intricacies of implementing large-scale machine learning in production with Sarah Aerni, Director of Data Science at Salesforce Einstein, in this insightful conference talk. Discover how Salesforce successfully integrates Agile methodologies into data science practices to serve over 100,000 customers. Learn about the company's innovative platform, including the open-source autoML library TransmogrifAI, and gain valuable insights into experimentation, deployment, and monitoring processes. Delve into the challenges of providing data scientists with effective tools for model deployment and continuous iteration. Understand the importance of rapid iteration, automated model retraining, and shipping billions of predictions daily. Explore strategies for detecting issues, identifying improvement opportunities, and maintaining a data science backlog through alerting and monitoring systems. Gain practical knowledge on fostering data science innovation within an Agile framework and learn from Salesforce's experiences in scaling AI applications across a vast customer base.

Syllabus

Introduction
Why are AI Machine Learning and Data Science still out of reach
What does it mean to move beyond giving your data scientists access
Salesforces approach to AI
Agenda
Building ML apps
No company is building one app
We need a third data scientist
Different degrees of skill set
Different data sizes
Classification
Language
Customization
Trust
Fixing leaks
Traditional AI process
Automation
Data Science Journey
Building Models
Getting Access to Data
Shipping Your App
Everyone Needs a Data Scientist
Data Scientists and Software Developers
Data Scientist
Building a Platform
Working Together
Finding opportunities for reuse
Transmogrify
Automated Pipeline
Data Sampling
Text Data
Stop Words
Learning Opportunities
Model Selection
The Job is Never Done
Metrics to Drive Agility
What Happens After Deployment
Minimum Viable Product
Agile Process
Agile Data Science
Monitoring
Model Monitoring
Investigate
Backlog
Focus
Key takeaways
Join the open source community
Thank you
Getting started in data science
ACM resources
Open source components
Platform secured experimentation
Latency considerations

Taught by

Association for Computing Machinery (ACM)

Reviews

Start your review of Agile Data Science - Achieving Salesforce-Scale Machine Learning in Production

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.