Data Science on Cloud Foundry
summary
Argues that Cloud Foundry removes ops friction so data teams spend time on models, not servers, and demonstrates concrete tooling: an Anaconda Python buildpack and an R buildpack, a deployed prediction API that stores models in Redis, accepts JSON for training/ingest/score, and does periodic retraining. Shows two live demos — a Transport for London disruption-duration predictor and an insurance customer-scoring app — and notes this works for small-to-medium data but large-scale training stays on HDFS/Spark or a Bosh‑provisioned cluster.
AI-written summary · about