Fulfilling God's purpose for my life
A disposable Bedrock Guardrails lab with Claude Haiku 4.5 and Converse: 34 prompts, what blocked, what slipped through, and a few false positives.
TypeSafe's Jev is a System One model: typed questions in, structured decisions out. I used it as a spend gate before ...
Nearly 20 years ago I asked God to hold my hand. A TechWorld with Nana talk about layoff, H1B, DevOps, and the people...
A small kagent demo on Akamai LKE: Alice can patch a broken Service, Bob cannot, and login alone is not enough.
How I automated my church's daily Bible verse video creation using Flask, Google Gemini, FFmpeg, Docker, and Railway ...
A small Kubernetes project I built to understand zero trust, and the three things I got wrong before I started to und...
Notes from an Amazon Bedrock AgentCore build: where prompt injection stops, where PII is masked, and when memory lags.
A Rust gRPC gateway for LLM traffic with policy checks, PII redaction, and metrics for OpenAI and Anthropic.
Add Claude Code Security to CI/CD as a second net, while keeping deterministic rules as the baseline.
Build a semantic cache with S3 Vectors, Bedrock, and Lambda so similar questions reuse prior LLM answers.
Design AI agent tools (MCP, LangChain, or custom) that cut token cost 60-70% without losing accuracy.
Build a security-gated CI/CD pipeline on AWS with secrets detection, SAST, CVE scanning, SBOM, and OPA on EKS and Far...
AWS DevOps Agent uses AI to investigate production incidents in seconds. A first look at what it does and where it fi...