Managing ML Lifecycles with Vertex AI with Erwin Huizenga
Play • 39 min

We’re learning all about Vertex AI this week as Carter Morgan and Jay Jenkins host guest Erwin Huizenga. He helps us understand what is meant by Asia Pacific and how Machine Learning is growing there. APAC’s Machine Learning scene is exciting for its enterprise companies leveraging ML for innovative projects at scale. The ML journey of many of these customers revealed challenges with things like efficiency that Vertex AI was built to solve.

The Vertex AI platform boasts tools that help with everything from the beginning stages of data collection to analysis, validation, transformation, model training, evaluation, serving the model, and metadata tracking. Erwin offers detailed examples of this pipeline process and describes how Feature Store helps clients manage their projects.

Using Vertex AI not only simplifies the initial development process but streamlines the iteration process as the model is adjusted over time. Pipelines offers automation options that help with this, Erwin explains. ML Operations are also built into Vertex AI to ensure everything is done in compliance with industry standards, even at scale. Using customer recommendations as an example, Erwin walks us through how Vertex AI can employ embedding to enhance customer experiences through ML.

By using Vertex AI in combination with other Google offerings like AutoML, companies can effectively build working ML projects without data science experience. We talk about the Vertex AI user interface and the other tools and APIS that are available there. Erwin tells us how Digits Financial uses Vertex AI and Pipeline to bring models to production in days rather than months, and how others can get started with Vertex AI, too.

Erwin Huizenga

Erwin Huizenga is a Data Scientist at Google specializing in TensorFLow, Python, and ML.

Cool things of the week
  • Announcing Spot Pods for GKE Autopilot—save on fault tolerant workloads blog
  • Indosat Ooredoo and Google Launch Strategic Partnership to Accelerate Digitalization Across SMBs and Enterprises in Indonesia site
  • Indosat Ooredoo dan Google Luncurkan Kemitraan Strategis untuk Percepatan Digitalisasi UMKM dan Perusahaan di Indonesia site
  • Vertex AI site
  • Google Cloud in Asia Pacific blog
  • Introduction to Vertex AI docs
  • What Is a Machine Learning Pipeline? site
  • TensorFlow site
  • PyTorch site
  • Vertex AI Feature Store docs
  • AutoML site
  • BigQuery ML site
  • Vertex AI Matching Engine docs
  • ScaNN site
  • Announcing ScaNN: Efficient Vector Similarity Search blog
  • Vertex AI Workbench site
  • Vertex Pipeline Case Study: Digits Financial site
  • Intro to Vertex Pipelines Codelab site
  • Vertex AI: Training and serving a custom model Codelab site
  • Vertex AI Workbench: Build an image classification model with transfer learning and the notebook executor Codelab site
  • APAC Best of Next 2021 site
  • TFX: A TensorFlow-Based Production-Scale Machine Learning Platform site
  • Rules of Machine Learning site
  • Google Cloud Skills Boost: Build and Deploy Machine Learning Solutions on Vertex AI site
  • Monitoring feature attributions: How Google saved one of the largest ML services in trouble blog
What’s something cool you’re working on?

Jay is working on APAC Best of Next and will be doing a session on sustainability!

Carter is working on transitioning the GCP Podcast to a video format!

More episodes
Clear search
Close search
Google apps
Main menu