Using AI to Land a Job Without Faking Your Way Through It.

There's a piece of advice circulating in every job-hunting thread, Slack group, and career subreddit right now. It goes like this: use AI as much as you want to prepare, then hide it during the actual hunt. Recruiters don't want to know. Polish your CV with it, draft your cover letters with it, rehearse with it — just never let anyone see the seams.

It sounds reasonable. It sounds safe. And for a while, it was.

It isn't anymore.

The reason is simple and a little uncomfortable: everyone is running that exact playbook. When every candidate uses the same tools to produce the same polished materials and then hides the fact, the outputs converge. Your AI-assisted CV looks identical to the next person's. And here's the part that should worry you — it also looks identical to the CV of someone who just cheated. On paper, responsible hidden use and outright faking are indistinguishable. The recruiter can't tell them apart, so the strategy stops paying off for the honest person running it.

This article argues for a different move, and it's aimed squarely at engineers with roughly two to six years of experience — the people who are good enough to do the job but junior enough that they still have to prove it on paper. The move is this: make your AI fluency legible. Not hidden, not pitched, not apologized for — legible, in the specific artifacts a recruiter already chooses to look at. Make it part of why you get hired instead of a thing you spend the whole process concealing.

Let me make the case, then give you the actual tactics.

Part I · Why the consensus brokeThe consensus stopped working, and here's why

The "use it but hide it" advice was forged in an earlier moment — when AI tooling was a genuine edge because not everyone had it, and when disclosing it carried real stigma because it was novel. In that world, quiet competence won. You did the work better and faster, said nothing, and let the results speak.

Three things broke that.

Put these together and you get the trap: the candidate who used AI responsibly — to sharpen their thinking, to prep harder, to close real gaps — produces an application that is visually identical to the candidate who used AI to fake competence they don't have. Concealment used to signal skill. Now it signals nothing, because it's what the cheater does too. You've worked hard to become invisible inside the slop.

The problem was never ethics. It's legibility. You need a way to be distinguishable from the faker — and hiding is the one move guaranteed to keep you looking exactly like them.

Part II · The wrong finish lineThe wrong endpoint

There's a deeper error buried in the hide-it playbook, and it's worth naming because it reframes everything that follows.

Hiding optimizes for the offer. It treats the moment of "you're hired" as the finish line. But that's the wrong endpoint. The right endpoint is week 12 — three months in, when the honeymoon is over and you're just doing the job.

Think about what each strategy actually produces at week 12.

The candidate who faked it — who presented an AI-augmented interview-self more capable than their real working-self — now has to sustain the illusion or get found out the first time the work demands more than a prompt can deliver. Concealment is brittle. It cracks under the load of real work. Every week is a small audit they might fail.

The candidate hired for their AI fluency has no such gap. The person who showed up to the interview is the person who shows up on day 1, who shows up at week 12. There's nothing to maintain because there was never a gap between the presented self and the working self. They were hired for how they actually work, so they get to just… work.

This is the core asymmetry, and it's the thing the consensus advice gets exactly backwards:

Faking it succeeds at the interview and fails at week 12. Hidden responsible use survives week 12 but loses the offer to the candidate whose workflow is legible.

Succeeding at the wrong test is worse than failing it, because it lands you in a job built on a version of yourself you can't sustain. Legibility is the only strategy that wins at both endpoints — you get the offer and you can do the job, because the offer was made to the real you.

So the goal isn't "disclose AI use because it's honest" (though it is). The goal is to make the durable, real version of how you work the thing that gets evaluated — so the job you land is one you can actually hold.

Part III · The tactical turnWhere legibility lives: artifacts, not pitches

Here's the most important tactical distinction in this whole piece, so I'll be blunt about it.

Legibility belongs in the artifacts a recruiter chooses to engage with — not in pitches you make during the screen.

The artifacts are your CV, your take-home submission, your portfolio. The recruiter decides which lines to read, which links to click, how deep to go. That's the key property: they opt in. A line on your CV that names your AI tooling sits there quietly. The recruiter who cares notices it and leans in. The recruiter who doesn't care skips it and reads the next line. You lose nothing either way.

Contrast that with pitching your AI fluency out loud in the screening call — "I want you to know I'm really good with AI tools." Now you've forced it into the conversation, made it a thing to be judged on the recruiter's terms in real time, and handed a skeptic the chance to disqualify you for the wrong reason before you've shown anything.

Artifacts protect you twice over. They protect against the recruiter who'd reject you reflexively, because the signal is passive and skippable rather than confrontational. And they self-filter for the right companies — a team that reads your "here's how I worked with AI on this" README and gets more interested is a team you want to work for. A team that recoils from it told you something valuable about the next three years of your life, and it cost you thirty minutes instead of three months.

One honest caveat before the tactics: yes, the hiring process is broken in places. Plenty of companies say they want AI fluency and then penalize any sign of it. Granted. But you're not here to fix hiring. You're here to optimize for the role you're about to live in — and legibility is how you make sure the role that's offered is one that fits how you actually work.

Now, the actionable part. I've split it into three buckets: prep that raises your floor invisibly, artifacts that carry the legibility signal, and how to behave in the room.

Prep that raises your floor (disclose nothing)

This first bucket is pure preparation. None of it gets mentioned to anyone. It's how you use AI to walk in better, not how you talk about AI. The point is to raise your performance floor so high that you're operating from competence, not luck, on the day.

Hostile STAR rehearsal. Most people rehearse their behavioral stories until they can say them once, smoothly. That's not enough — interviewers probe. Have the AI play a hostile, skeptical interviewer and drill each story three "why" levels deep. You say your story; it asks the follow-up; you answer; it asks the follow-up to that. Keep going until your stories survive the second and third follow-up, not just the first. Most candidates fall apart at the second "why." If yours hold, you're already ahead.

Stack-specific mocks, not generic LeetCode. Generic algorithm grinding has diminishing returns past a point. Far more valuable: mock interviews on the company's actual terrain. Feed the AI the job description, the languages and frameworks they list, the kind of data model or domain they work in, and have it generate practice problems and system-design prompts that match. If they're a payments company on Go and Postgres, you practice payments-shaped problems in Go against a relational schema — not abstract tree-traversal puzzles. You want to arrive fluent in their world.

Interviewer recon. If you know who's interviewing you, have the AI read their conference talks, their GitHub, their blog posts, and synthesize what they care about. The goal isn't to flatter them. It's to walk in with one genuinely engaged, opinionated question or observation that shows you've thought about the space the way they have. There's a world of difference between a question that sounds googled and one that sounds like you've actually wrestled with the same problem they have.

Gap-close. This is the most underused one. Have the AI diff your real experience against the job description and name the gaps honestly. Then close the smallest credible gap — the one you can actually close in a weekend project or by genuinely reading one foundational paper. The point isn't to fake the gap away. It's so that when the gap comes up in the interview, you can answer it honestly: "I hadn't used Kafka in production, so last weekend I built a small pipeline with it to understand the failure modes — here's what surprised me." That answer beats both dodging the question and pretending the gap doesn't exist. You turned a weakness into evidence of how fast you close gaps.

Notice what all four have in common: they make the real you better. Nothing here is a costume. You walk in genuinely more prepared, and none of it is anything you'd ever need to disclose, because it's just preparation — the same category as practicing in a mirror, only sharper.

Artifacts that carry the signal

This is where legibility actually happens. These are the places where you let your AI fluency show — quietly, skippably, on the recruiter's terms.

The CV stack-line. Somewhere in your skills or tooling section, alongside your languages and frameworks, name your AI tools as part of your stack. Not a paragraph, not a manifesto — a line. The way you'd list Docker or Postgres. Something like: Tooling: Git, Docker, Postgres, Claude Code / Cursor for development workflows.

The recruiter who cares about AI fluency notices it and reads you as someone who knows what these tools are for. The recruiter who doesn't care skips the line entirely, exactly as they'd skip past a tool they don't recognize. The asymmetry is entirely in your favor: upside if they care, nothing lost if they don't.

The take-home "tools used" note. Only if AI isn't explicitly forbidden by the take-home's instructions — read them carefully, and respect them if it is. When it's allowed, add a short paragraph to your submission's README: what you used AI for, and what you deliberately didn't.

This single paragraph reframes you completely. Without it, a polished submission invites the suspicion "did they even write this?" With it, you've shown the actually-valuable senior skill: knowing what to delegate and what to own. You've moved from possibly cheated to knows how to work.

One portfolio piece where the prompt chain is the artifact. Build or document one project where the interesting thing isn't just the output but how you got there with AI. The README explains your prompt chain — and crucially, why this prompt and not that one. Where did the AI go wrong and how did you catch it? What did you reject? Which decision did you overrule it on? This is the single most differentiating artifact you can produce, because it demonstrates judgment — the thing that's actually scarce. Anyone can get output from a model. Far fewer can show a legible reasoning trail for steering one. That trail is the proof that you're the operator, not the operated.

In the room: incidental, not a pitch

When AI does come up in conversation — and it will — the entire trick is tone. Mention it the way you'd mention any other tool. Unremarkably.

Don't reframe it as a special skill. Don't ask the interviewer to evaluate you differently. Don't request that any rules be changed. Just refer to it in passing, the way you'd refer to your IDE or your terminal.

The treating-it-as-unremarkable is the signal. It tells the interviewer you're someone for whom AI is a normal, integrated part of how engineering gets done — not a crutch you're nervous about, not a trophy you're showing off. Confident incidental mention communicates fluency far better than any pitch could, precisely because it doesn't try to.

The deeper logic: a pitch makes AI use the subject of evaluation. An incidental mention makes your engineering judgment the subject, with AI as one ordinary tool serving it. The second framing is the one you want, every time.

Part IV · The objectionThe objection, answered

The strongest argument against everything above is real, so let's meet it head-on: "Some companies will screen you out the moment they see you naming AI. You're handing skeptics a reason to reject you."

True. Some will. And here's why that's a feature, not a bug.

If a company would disqualify you for being legible about how you actually work, then how you actually work is incompatible with that company. You were going to find that out eventually — the only question was whether you'd find it out in the thirty minutes of a screening process or in the three months of a job that slowly reveals it doesn't fit. The legible move surfaces the mismatch early, when it's cheap, instead of late, when you've already quit your last job and signed the offer.

So the legible move isn't a risk you're taking. It's a fit filter you're running.

Every company it screens you out of is a company that just saved you from a bad three years. And every company it draws toward you is one that wants the engineer you actually are — which is the only kind of offer worth taking.

So, concretely

If you're two to six years in and about to start hunting, here's the whole thing in one breath:

Use AI hard in prep — hostile STAR drills, stack-specific mocks, interviewer recon, honest gap-closing — and disclose none of it, because it's just preparation. Then let your fluency show legibly in the artifacts the recruiter opts into: a stack-line on the CV, a tools-used note on the take-home, one portfolio piece where the prompt chain and your steering of it is the work. In the room, mention AI the way you'd mention your editor — incidentally, never as a pitch. And when a company recoils, thank it for the early filter and move on.

You're not trying to beat the hiring process or game it. You're trying to land in a role you can still do well at week 12 — as the real, AI-fluent version of yourself, not a version you have to keep hiding. That's the job worth getting. Legibility is how you get it.