
Vertex AI Search vs RAG Engine: Grounding Gemini with My Own Data
Comparing three grounding methods on Google Cloud using my own blog posts. Vertex AI Search, RAG Engine, and Google Search grounding. What worked, what broke, and what surprised me.

Comparing three grounding methods on Google Cloud using my own blog posts. Vertex AI Search, RAG Engine, and Google Search grounding. What worked, what broke, and what surprised me.

A bash script that downloads audio from YouTube, uploads it to S3, runs an Amazon Transcribe job, and outputs a plain text transcript.

After 40+ AWS lessons and 7 certifications, I'm creating my first Google Cloud course. This post documents everything I learned setting up my Google Cloud environment, from account creation to IAM configuration, through the lens of an AWS engineer discovering a new cloud.

How to track project costs for Bedrock Agents using AWS Budgets. Covers the limitation with on-demand model tagging and the solution using Application Inference Profiles.

Learn what makes AI systems agentic and how to design autonomous AI architectures on AWS using Amazon Bedrock, Lambda, Step Functions, and persistent memory services.

Protect public API Gateway endpoints from abuse using usage plans and budget alerts. This post covers practical strategies to limit requests, monitor costs, and explores when AWS WAF or authentication might be needed.

After launching my first Bedrock agent, I wanted to add proper observability. This post documents how I added API Gateway access logs, CloudWatch Logs Insights, and DynamoDB for conversation storage to debug issues, monitor performance, and understand usage patterns.

Build an Amazon Bedrock agent that queries the AWS News Blog RSS feed and answers questions about recent AWS announcements. Use Lambda action groups to fetch and parse the feed, expose the agent through Amazon API Gateway, and provide a terminal-style web UI.

A workflow for recording AWS Console demos and exporting 4K video for YouTube. Compares Dell 4K and Apple Studio Display setups.

This is the story of an agent. What began as a simple Amazon Bedrock agent demo evolved into a real, production agent running on my own website. It connects my documentation, notes, and projects into a single conversational interface. This started as a small experiment. Then it became a website feature. Now it feels like something bigger. v1.0 is just the beginning.