I'm the person teams call when their AI outputs are technically correct but somehow still terrible. My job is making LLMs actually useful, not just impressive in a demo, but reliable when thousands of real users are on the other end. At Regie.ai, I engineer multi-step prompts for sales outreach workflows: data extraction, buyer signal filtering, personalized cold emails, the whole funnel. Prompts for profit, basically. I took first-round relevance from a rough 52% to over 85%, evaluated 4,500+ AI outputs (yes, I read them all), and turned a 3-day QA process into a 1-day one because nobody has time for that. Before I fell into the AI rabbit hole, I was a content person. Wrote 400+ long-form articles at Headout, one hit 1.4M impressions, which still feels unreal. I also designed how 5 chatbots should "sound" and built 300+ canned responses that cut support reply times by 96%. Turns out, years of obsessing over word choice is surprisingly useful when you're trying to make AI not sound like a robot. What I'm good at: prompt engineering that scales, conversational AI that doesn't make users want to scream, and building systems so the work doesn't fall apart when I'm not looking. What I'm not: a backend engineer. But I speak their language well enough to debug together without anyone crying. If you're building something with AI and need someone who cares about the details , let's chat.…I'm the person teams call when their AI outputs are technically correct but somehow still terrible. My job is making LLMs actually useful, not just impressive in a demo, but reliable when thousands of real users are on the other end. At Regie.ai, I engineer multi-step prompts for sales outreach workflows: data extraction, buyer signal filtering, personalized cold emails, the whole funnel. Prompts for profit, basically. I took first-round relevance from a rough 52% to over 85%, evaluated 4,500+ AI outputs (yes, I read them all), and turned a 3-day QA process into a 1-day one because nobody has time for that. Before I fell into the AI rabbit hole, I was a content person. Wrote 400+ long-form articles at Headout, one hit 1.4M impressions, which still feels unreal. I also designed how 5 chatbots should "sound" and built 300+ canned responses that cut support reply times by 96%. Turns out, years of obsessing over word choice is surprisingly useful when you're trying to make AI not sound like a robot. What I'm good at: prompt engineering that scales, conversational AI that doesn't make users want to scream, and building systems so the work doesn't fall apart when I'm not looking. What I'm not: a backend engineer. But I speak their language well enough to debug together without anyone crying. If you're building something with AI and need someone who cares about the details , let's chat.WWWW…