Digital Product

The CV Operating System

by annapoorna

The AI-powered CV system that got me interviews in London

The CV Operating System

What you'll learn

why your CV isn't getting responses

and why it's probably not your experience. The ATS filter, the one-CV trap, and the language mismatch: what's actually sitting between your application and a human.

how to write bullets that pass both the filter and the human

the constraint system: what every bullet must have, what none of them can, and how to tell the difference between a strong signal and a task description.

how to define your positioning before you write a word

the single step most people skip. How to articulate which version of you a specific role is hiring, and why

how to know when your CV is actually working

the judgement layer. How to read your own CV the way a recruiter does, spot the common failure modes, and know when to stop iterating.

how to use AI without producing generic output

the exact prompts, in sequence, that give AI the direction and constraints it needs to write in your voice, not everyone else's.

how the same experience becomes three different CVs

the three-CV proof shows the same 10-year career positioned for three different roles. Same facts, different story entirely.

The AI-powered CV system that got me interviews in London that works for any geography.

Frequently asked questions

This is a decision system, not a template. Anyone can Google a CV template. What this gives you is a repeatable process — how to position yourself before you write a word, how to use AI without producing generic output, how to pass the ATS filter, and how to know when your CV is actually working. It took me one job hunt, hundreds of applications, and months of iteration to build this. You're buying the shortcut to what actually works — not the journey of figuring it out the hard way.

You can. And you'll get a polished-sounding CV that sounds like everyone else's. The problem is not the AI — it's the input. Garbage in, polished garbage out. This guide gives you the exact prompts, the positioning logic, and the constraint system that stops AI from producing generic output. The value is in the thinking layer before you open the tool — that's what this document teaches.

Most guides give you advice. This gives you a workflow you open alongside your CV and follow step by step — fit assessment, positioning hypothesis, section-by-section prompts, a constraint checklist, a final prune prompt. It's designed to be used, not read once and shelved. If you're actively job hunting right now, you'll use it on your next application.

There is a skeleton of a template at the end — but that's 10% of it. The bulk of the document is the decision system: how to define your positioning before you write, which stories to select and why, how to write bullets that pass both ATS and human scrutiny, and how to iterate until it's actually sharp. The template tells you where to put things. The system tells you what to put there.

It's a PDF. You can open it on your phone, laptop, or tablet — no special software needed. The prompts are copy-paste ready for ChatGPT, Claude, or any AI tool you already use.

Yes. Part Eight shows the same 10-year career written three completely different ways — targeting a transformation role, a BizOps role, and an in-house strategy role. Same experience, three different CVs. Reading them side by side is where the positioning logic clicks. It's the most useful part for a lot of people.

Yes. The system was built on consulting / strategy experience but the mechanics are universal — ATS filters don't care what your background is, and neither does the positioning logic. A lawyer moving from litigation to in-house has a positioning problem. A PM moving from early-stage to enterprise has a language problem. An engineer moving into product has a framing problem. This system solves all three. The prompts, the constraint rules, the iteration loop — none of it requires an MBA. It requires that you have experience worth framing well. If that's you, this works.

Specifically yes. I wrote this after a 3-year sabbatical and a full job hunt with a non-linear career behind me. The guide includes a specific framing for career breaks — name the period, frame it in one professional clause, move on. There is also an example positioning hypothesis written for re-entry after a career break. This was not written for someone whose career is a clean straight line.

The core of this — positioning logic, ATS rules, bullet discipline, iteration process — works anywhere. The examples and framing are calibrated for the UK market (I job-hunted here and got offers here), but the system is not UK-specific. If you're applying in a market where ATS is standard and English CVs are the norm, this works.

The system was designed for exactly you. The problem it solves — how to compress 10+ years into 2 pages without losing signal, how to make the same experience read differently for different roles — is a mid-career problem, not an entry-level one (see "I'm a student / fresh grad" above if that's you). The three-CV proof in Part Eight uses a 10-year career as its base. The more experience you have, the more you need a decision system, not just a template.

Yes — and there's a section specifically for this. The core insight: don't try to prove you know the new industry. Lead with transferable capability that's rare there. The positioning hypothesis becomes: X capability, now applied to Y sector. That's the frame that actually lands.

Yes. Start with the Fit Assessment prompt in Part Four — paste your existing CV and the JD and ask AI to assess the fit, identify the gaps, and suggest the strongest angles. That single step will tell you exactly what needs changing and what's already working. You don't have to rebuild from scratch unless the positioning is fundamentally off.

honestly: the positioning mechanic needs a few different possible framings of a career to choose between, which a thin CV usually doesn't have yet; the ATS rules, constraint system and iteration loop still apply broadly; if unsure, message first