Blog
The Bridge.
AI systems, sales engineering, infrastructure, and the boring parts that matter. Writing from the intersection of technology and business.
From Support Desk to Solutions Consultant: My Certification Journey
My career path wasn't a straight line — it was a deliberate build, certification by certification, skill by skill. Here's the roadmap that worked for me. The Foundation: IT Service Management (2012-2014) I started with the HDI certifications (KCS, HDM, SCTL) — these taught me that support is a strategic function, not a cost center. … Read more
Read post →Ratnarok: Building an AI-Driven Tabletop RPG
What happens when you combine AI with tabletop RPGs? You get Ratnarok — a web-based RPG where players are heroic rats in a post-human world, powered by AI-driven narrative. Three Game Modes Solo Mode — Fully procedural AI-generated adventures with Gemini driving NPC dialogue, quests, and world events Semi-Guided — AI + optional human Game … Read more
Read post →AI-Powered Attendance Intelligence: OCR with Vision LLMs
Schools manually transcribe attendance sheets every day. It's slow, error-prone, and consumes hours of administrative work. Traditional OCR tools struggle with handwritten or varied-format sheets. Enter Presença OCR. The Solution A Python/Flask backend using GLM-4 Vision LLM (ZhipuAI) processes scanned attendance sheets. Teachers upload photos and get structured data instantly — no manual transcription needed. … Read more
Read post →Running 15+ Production Services on a Notebook: My HomeLab Setup
My HomeLab runs on a single notebook 24/7 — and it powers 15+ production services including this very blog. Here's how it all works. The Stack Hypervisor: Proxmox VE for VM management Containers: Docker Compose + Docker Swarm Networking: Cloudflare Tunnel (zero open ports) + Tailscale VPN DNS: Terraform-managed Cloudflare DNS Observability: Prometheus + Grafana … Read more
Read post →Portfolio Management with AI: Building a Strategic Scoring Engine
As a creator with multiple projects running simultaneously, I faced a constant challenge: what should I work on next? Gut feel wasn't cutting it. I needed data. The Strategic Value Score (SVS) PorfolioManager calculates a multi-dimensional score for every project based on: Strategic alignment — how well does this project serve my long-term goals? Market … Read more
Read post →The Spec-Driven Development Pipeline: A Methodology for AI-Assisted Coding
AI coding agents are incredibly powerful, but they produce inconsistent results when given vague instructions. After months of working with OpenCode, Cursor, and Claude Code, I developed a structured methodology: the SDD Pipeline. The 7 Stages The pipeline transforms a raw idea into executable AI prompts through sequential specification stages: Idea — One-sentence concept capture … Read more
Read post →Offline Speech-to-Text: Why I Built ditado Instead of a Subscription
I talk to AI agents all day. Typing responses breaks flow. Existing dictation tools either require a subscription, send your audio to the cloud, or both. So I built ditado — an offline Windows dictation app that never sends your data anywhere. How It Works Press F8 to start recording, press F8 again to stop. … Read more
Read post →From Local Script to Multi-Tenant SaaS: The JobMiner Story
Every SaaS starts somewhere. JobMiner began as JobHuntingCRM — a local Flask app with SQLite that I built for my own job search. It worked so well that I decided to turn it into a proper multi-tenant SaaS platform. The Evolution Phase 1: JobHuntingCRM — A local Python Flask app with Gemini AI integration. It … Read more
Read post →Self-Hosted AI Observability: Why Every AI Agent Needs a Trace
When you're building autonomous AI agents, the scariest question is: what did it actually do? Without observability, every agent session is a black box — you see the input and the output, but everything in between is guesswork. The Observability Stack I deployed a full OpenTelemetry stack on my HomeLab to trace every AI agent … Read more
Read post →Building a Multi-Agent AI Orchestrator: Lessons from CourseSmith
When I set out to build CourseSmith, I had a simple thesis: if AI can generate code, it should be able to generate structured educational content. The reality was far more complex — and far more interesting. The Problem with Single-LLM Content Generation Early experiments with single-prompt course generation produced shallow results. A single LLM … Read more
Read post →