85% of Machine Learning projects fail (according to Gartner) and we want to talk about pivotal decisions you need to make that will make or break your ML project. The goal of this talk is to ignite conversation about delivering solutions to customers' problems in a world of imperfect data.
In the last 4 years we've been working on extracting value from the data OpenShift clusters stream. One of our projects goes live around DevConf (June 2023): predicting if an OpenShift cluster will have issues during its next upgrade and giving admins advise on what needs to be addressed.
The lessons we've learned translate to many projects, and we wish we didn't have to learn those lessons the hard way: - keeping business goal in mind and always talking to your stakeholders (business and customer-facing associates) - defining success metric early - not blindly following latest coolest technologies if they produce equal or worse results than an architecturally more simple approach
Wearing many hats including a red fedora: software engineer, data scientist and data engineer, using data to improve products and make people more productive.
Create a custom schedule.
Take it with you on mobile.
Get listed in the directory.
