How We Built A Scalable Serverless Machine Learning Pipeline For Under $7/Day

A few months ago, Adam Tankanow, principal engineer at CloudZero, read a blog by Andreesan Horowitz about how any company using  ML or AI is destined to have weaker margins and higher costs. He disagreed.

In this talk, Adam will discuss why machine learning doesn't need to cost a fortune, and how he and his team successfully built a machine learning pipeline with cost constraints that help maintain strong margins. 

In this session, Adam covers: 

• A step by step process engineering teams can take to productize their work before shipping it. 
• How his years writing code for audio equipment taught him a process you can apply to any engineering process to ensure you're not tackling expensive tech debt down the road.
• How CloudZero used a novel approach of working backward from a $7 a day operational cost requirement to build a scalable serverless machine learning pipeline 

• All listener-submitted questions.

About Adam:

Adam Tankanow is a Principal Software Engineer and member of the CloudZero founding team. Prior to CloudZero, Adam worked as an audio engineer and managed a pre-cloud distributed speech recognition farm. He is an AWS Professional Certified Solutions Architect.