![](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgWR4oChL3nihotk5T-dSWtYWs7fTew6jnIQ9_KtrVC2j8oloAOjvaXNslyr8fbld5NQ2yVlLpWocgEA5kNE0zDJFoN4NqsPkI1U3M9uAZCTVzN06UXmnhImKFkex-Jrq-FIGYouWwJ1f1ZhsHPNK-2b4o0gzbJjSKbYRkfMRVbBdGJgrOnCVGOaA/s200/20241205_171946_GoogleAI.jpg)
![](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhSnj9ArcLsIAFtDtS2eIEtYH_lE3PWNVZxVn3-_UL02oRueI2Y0uT9K3wTvuaQykuCmkwYmovu-9-uAET5aAlWrPvgTr0Y89oDRho2-RH4DwwUxGb4qAm9ONu6_6U7chsTcafhcM-mEBsZYnwdPJ8RtGPySIGtUuebcOBhkvevIGXjg1dior7aPA/s200/20241205_171839_GoogleAI.jpg)
I arrived a few minutes before 1:30pm, just before the lab supposed to start. Surprised to see so few attendees (6? 7?), and out of them, I knew 2. They are organizers of GDG Seattle. Seems the goal today is introducing AlloyDB AI and Vertex AI Agent Builder to create an AI agent for a shopping chatbot. Reference: https://github.com/GoogleCloudPlatform/generative-ai
First lab is using simple SQL in Google's AlloyDB to set up tables and embedding. However, for these training accounts, it times out so quickly, that we could only load 8 rows of data. 2nd lab is building the AI agent using the model setup in lab #1. If you are not familiar with Google Cloud, it's hard to find where things are. There were 3 Google employees (other than the 2 teaching the labs) onsite to troubleshooting our errors (usually caused by missing steps of the labs or things changed since the labs were tested). We ended with a quiz and a survey. Everyone got a pair of very colorful socks. (why socks?)
![](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQGvnX4DI75E7nkx9ZfDXkR5wX5y_yNOqh_4JYwL4y-aXZcdEUl1gCWNNf1qRzeN9_UZ_mI31jLsebpTJL58YKGmTQfVo9TBPJgoluF2nlq6zJSVqC5hOBtEq-mtjbJuT8b3xtljafSCRP5xhbqMUlsjfV7GCJq1Q64OdQLnCfFXAQxbyB0e7uew/s200/20241205_171851_GoogleAI.jpg)
One cool thing is that 2 artists were hired to draw cartoons of us. I received 2 very colorful drawings. They are very fast.