Building AI Agents, Breaking Code, and the Quest for "Zero-Slop" Engineering

Hey everyone. I'm finally jumping into the DEV community to share what I'm learning while building in the current AI-agent gold rush.

What I'm Building

I'm a hobbyist builder obsessed with AI agents and automation. When I'm not in my day job, I'm usually elbow-deep in a few different side projects:

• AI Agent Infrastructure: Experimenting with how to make agents actually reliable (spoiler: it's harder than the Twitter demos suggest).
• Phishing Detection Engines: Building out a system called HydraG to see if we can out-engineer the latest wave of AI-driven security threats.
• Voice AI Experiments: Messing around with real-time voice receptionists and the latency/reliability trade-offs involved there. My toolkit of choice lately is Node.js, Python, Supabase, n8n, and Lovable.dev. I'm a big believer in shipping fast but shipping clean—which is becoming a lost art in the age of AI-generated code.

My Engineering Philosophy

I’ve developed a bit of a "Root Cause Mandate." I noticed early on that AI tools love to patch symptoms instead of fixing the actual problem. I’m interested in the friction points between us and the machines:

• Why AI tends to stop at the first "good enough" answer.
• How to implement mandatory critic audits in your build pipeline.
• Designing for "zero-slop"—keeping the codebase lean and intentional even when an LLM is doing the heavy lifting. Why I'm Here

I'm here to document the "war stories" from my projects. No fluff, no 2000-word SEO intros—just the technical hurdles I'm hitting and the patterns I'm using to solve them.

If you’re building agents, wrestling with LLM behavior in production, or just love a good debugging deep-dive, let’s connect.