Pies, Lies and AIs

Exploring the world of data and organisational intransigence

1. AI doesn't have to be expensive

Recently we looked at a project to predict train delays on the network. It's a complex problem and we addressed a small part of it. It was a real collaboration effort between real-time engineering and integration and data flow with predictions. My data science teams prefer to use either #ApacheSpark or #Python (sci-kit learn or #keras normally). In this we ruled out #Spark (at least #Databricks) as we originally wanted to get the entire solution under a budget of £100. That w

 

Subscribe Form

Stay up to date

 

CONTACT

Thanks for submitting!