Hey, I’m Zed Haque — a voice agent specialist who spends way too much time thinking about how humans and AI can have natural conversations.

I work primarily with OpenAI’s Realtime API, building voice agents that actually work in production. Not demos, not prototypes — real systems that handle real users and real edge cases. The kind of stuff that breaks at 2 AM and teaches you lessons you won’t find in documentation.

What I Do

• Build production voice agents using OpenAI Realtime API
• Write detailed technical guides based on real implementations
• Share practical patterns for WebRTC, WebSocket, and audio streaming
• Help developers avoid the mistakes I’ve already made
• Document the messy reality of voice AI development
• Create practical frameworks and templates for entrepreneurs — check out my Disciplined Entrepreneurship Notion template for systematic startup building

Why This Site?

The voice AI space moves fast, and documentation rarely keeps up with reality. I started writing to fill that gap — capturing the practical knowledge you need when you’re actually building something.

No marketing fluff. No “revolutionary” claims. Just honest technical content about what works, what doesn’t, and why. I’ve published 56+ articles covering everything from basic WebRTC connections to advanced function calling patterns.

The Approach

I believe the best way to learn is by doing, and the best way to teach is by showing real implementations. Every article on this blog comes from hands-on experience — problems I’ve solved, patterns I’ve discovered, and mistakes I’ve made (and hopefully won’t make again).

If you’re building voice agents, dealing with latency issues, wrestling with WebRTC connections, or just trying to figure out why your function calling isn’t working, this blog is for you. Welcome to the messy, fascinating world of voice AI development.