deeflect

Tag

ai-engineering

7 posts

The 3-Month Gap: Building AI Agents That Actually Work

3 weeks ago

The hard part of building AI agents isn't the demo - it's the 3 months after. Real lessons from running a 15-agent system in production for 90 days.

9 min

Universal MCP Server: Two Tools, 56 APIs

1 month ago

How I built a universal MCP server that wraps 56 APIs into just two tools using the OpenAPI Code Mode pattern - cutting token costs 50x.

11 min

Prompt engineering is dead

2 months ago

Prompt engineering had an 18-month run as a real skill. Models killed it. Learn what actually matters now - tool use, agent architecture, and MCP integration.

8 min

My coding stack is 4 models deep

6 months ago

How I use Grok, Claude Code, Codex, and more in sequence to ship faster - what each model does best and why one tool isn't enough.

8 min

Why AI Products Have Terrible UX

7 months ago

Most AI products have terrible UX - not because the AI is bad, but because no one who understands both AI and design is building them. Here's what to fix.

9 min

How ADHD and AI Work Together (My Real Stack)

8 months ago

Scored 147 on RAADS-R and 38 on AQ-50 (autism), plus ADHD. Here's how I built AI systems that work with my brain, not against it. Real stack, real workflows.

9 min

Why I Left My Role as Solo Product Designer at VALK

8 months ago

I was the only designer at a $4B fintech platform for 5 years. Here's the real reason I left in 2025 and what I'm building now.

7 min