@MazlumTosun3@GoogleCloudTech It does remind me the integration testing framework I’ve created at L Oreal that I presented and where you were present and asked many questions. Especially the given when then framework. Shit ❤️
A huge part of the #ML process is experimentation and tuning. Luckily, there are a few Vertex AI features that can help you with tuning and scaling your ML models.
Press ▶️ to learn how you can get models out of experimentation and into production with #VertexAI ↓
Create Kubernetes clusters anywhere, via the cloud? We're just about there, with the public preview of bare metal provisioning in the @googlecloud Console from @fonteny and team.
Docs: https://t.co/9b5NnrGvUw
Try it using GCE: https://t.co/BZNhCgYfh3
@Francois_Nguyen I’m fan of number 3 « Metrics layers will unify the data stack » and number 8 « Data Observability becomes a Must Have » I see these subjects becoming mainstream in the near future
For intro data science this semester we’ve been using #DuckDB in Python notebooks as a way to manipulate pandas data frames, and the students like it SO MUCH MORE than the equivalent pandas syntax
Object tables (a new table type in #BigQuery) enables you to directly run analytics and #ML on images, audio, documents and other file types using existing frameworks like SQL and remote functions natively in BigQuery itself.
Learn more → https://t.co/mfXa4XPs9h
Object tables (a new table type in #BigQuery) enables you to directly run analytics and #ML on images, audio, documents and other file types using existing frameworks like SQL and remote functions natively in BigQuery itself.
Learn more → https://t.co/mfXa4XPs9h
The greatest trick the cloud vendors + Snowflake ever played was convincing orgs that by spending more on storage (of raw data) and compute (data modeling), teams can transform garbage data into valuable insights