One app and a plan I have no idea will work

cv-tailor started as a personal problem. You build your profile once. The agent runs daily, finds relevant listings, scores them against your profile using AI, and sends you a digest. It learns from what you save and discard — building a preference model over time so the scoring gets better the more you use it. You also get a clean ATS-proof CV generated from your profile in multiple styles.

I built it because the problem was real and nobody had solved it the way I wanted.

I did not plan to build all of this. I built what I needed, then kept going because each problem revealed the next one.

I am an Engineering People Manager. I am good at building systems, running processes, managing people, and shipping internal tools. I have never sold anything. I have never found customers. I have never taken something from zero to revenue. That whole side of building a product — the market, the positioning, the distribution — I had no exposure to it and no instinct for it.

So I did what I do when I do not know something. I read, I researched, talked to people, and built a framework. Seven stages from zero to your first ten paying customers. Each stage with entry criteria, specific steps, and exit criteria you have to meet before moving on. Then I built GTM OS — the framework turned into a working tool that walks you through each stage, forces the right questions before you move forward, and keeps you honest about where you actually are versus where you think you are.

I am going to apply it to cv-tailor in real time and write about what happens. Not because I think it is right — I genuinely do not know if it is. But because testing it on a real product in a real market is the only way to find out.

I build with Claude, Claude Code, Gemini, GLM, Kimi, Qwen, and whatever model is worth trying this week. I switch often and adapt fast. That is also part of what I am learning.

I am not writing this to build a personal brand or position myself as an expert. I am writing it because building in public feels like the honest way to do this — and because if any of it works, this will be the story of how it happened.

By the end of this year I want one thing to be true: cv-tailor has at least one paying user who is not me.

I am a beginner at this. Building products people pay for is not something I have done before. That is exactly why I am writing it down.

This blog publishes weekly. One post, one topic, what actually happened.