Insights on AI, Cloud
& Modern Engineering
We write about AI agents, cloud architecture, cost optimization, and the tools we use every day to build software.
Gemma 4 + MCP + AWS: Build Self-Hosted Agentic AI with Function Calling & Tool Use
Complete guide to connecting Gemma 4's native function calling to MCP servers on AWS. Covers the gemma-mcp package, custom MCP server development, EC2/SageMaker/Bedrock AgentCore deployment, multi-tool agent architecture, and production cost optimization.
Gemma 4 vs Llama 4 vs Qwen 3.5: Open-Weight Model Comparison for Production in 2026
Three open-weight model families, one decision. We compare Gemma 4, Llama 4, and Qwen 3.5 across benchmarks, licensing, inference speed, memory, multimodal capabilities, and production readiness with a clear decision framework.
How to Fine-Tune Gemma 4 with LoRA & QLoRA: Complete Guide for Custom Models
Fine-tune any Gemma 4 model (E2B to 31B) with LoRA and QLoRA using Hugging Face, Unsloth, or Keras. Covers dataset prep, hyperparameter tuning, evaluation, and deploying your custom model to production.
Deploy Gemma 4 on AWS: EC2, SageMaker & Inferentia Cost Comparison Guide
Production-ready Gemma 4 deployment on AWS. We compare EC2 GPU instances, SageMaker endpoints, and Inferentia2 chips with real cost breakdowns, auto-scaling strategies, and optimization tips for each Gemma 4 model size.
Build an AI Agent with Gemma 4: Function Calling, Tool Use & MCP Integration
Gemma 4 ships with native function calling via 6 special tokens. We show how to build a production AI agent with structured tool use, MCP server integration, multi-step reasoning chains, and real-world agentic workflows.
Gemma 4 on Edge: Running Multimodal AI on Mobile, Raspberry Pi & IoT Devices
Gemma 4 E2B and E4B bring multimodal AI (text, image, audio) to phones and edge devices with 128K context. We cover on-device deployment with MediaPipe, LiteRT, MLX, quantization strategies, and real-world latency benchmarks.
Using OpenClaw with Gemma 4: Complete Local AI Agent Setup Guide
Run OpenClaw with Google's Gemma 4 via Ollama for a fully self-hosted, zero-cost AI agent. Covers model selection, configuration, function calling, custom skills, and performance tuning.
How to Develop an EmDash Plugin: Local Environment Setup, Testing & Deployment Guide
Step-by-step guide to building EmDash (Cloudflare) plugins from scratch. Covers local environment setup with SQLite/workerd, plugin scaffolding, capability manifests, lifecycle hooks, testing with miniflare, debugging, and deploying to Cloudflare Workers.
Google Gemma 4 Developer Guide: Benchmarks, Architecture & Local Deployment
Google DeepMind's Gemma 4 ships 4 open-weight models (2.3B–31B) under Apache 2.0 with 256K context, native multimodal, and function calling. Full benchmark breakdown, architecture deep-dive, and local setup guide.
Cloudflare EmDash Developer Guide: Setup, Plugins, Themes & Deployment
Everything you need to get started with EmDash, Cloudflare's open-source TypeScript CMS. Covers installation, sandboxed plugin development, Astro 6 theming, WordPress migration, and deployment on Cloudflare Workers or Node.js.
Migrating WordPress to EmDash: Complete Step-by-Step Guide
Step-by-step guide to migrating your WordPress site to Cloudflare's EmDash CMS. Covers WXR export, EmDash Exporter plugin, custom post type mapping, media migration, URL preservation, and theme rebuilding in Astro 6.
EmDash vs WordPress vs MedusaJS: CMS & E-Commerce Platform Comparison 2026
Honest comparison of EmDash, WordPress, and MedusaJS across architecture, security, hosting costs, developer experience, and ideal use cases. Real pricing numbers and trade-offs from a team that builds on all three.