AI for job search: tailor applications and practise interviews with Llama

A job search is mostly reading and writing: studying ads, tailoring a CV, writing letters, preparing answers. Llama helps with all of it, quickly and for free. What it cannot do is know the employer, today's job market or current pay in your city, because it has no internet access and its knowledge stops at a cut-off date. Use it to sharpen what you already have, and never to invent experience you do not have.

Step 1: decode the job ad

Start with the employer's side. Most ads mix true must-haves with wish-list items, and knowing which is which tells you what to put first.

Here is a job ad. List: 1) the five most important requirements, in order of how much the ad stresses them; 2) skills it mentions only indirectly; 3) questions I should be ready to answer in an interview; 4) anything unclear I should ask the recruiter. Job ad: [paste]

Open this prompt in the chat and paste the ad you are looking at.

Step 2: tailor your CV, honestly

Paste the ad and your current CV, with your phone number, address and other personal details removed. Then ask for specific, checkable suggestions rather than a full rewrite:

Weak:
Improve my resume.

Strong:
Read my CV next to the job ad below. For each of the top requirements,
tell me which line of my CV supports it, or write "not shown". Then
suggest rewrites for up to 6 bullet points so they start with an
action verb and show a result. Use only facts that are in my CV.
Where a number would help, write [ADD NUMBER] instead of inventing one.

That last instruction matters. Asked to "make it impressive", models readily add percentages and achievements that never happened. Interviewers will ask about them, so every claim must be true and something you can talk about with confidence.

Step 3: a cover letter in your own voice

Recruiters read many letters that could have been sent by anyone. Yours stands out when it contains things only you could write, so give Llama that raw material:

• Why this company or this role, in your own words, even if they are rough.
• One story where you solved a problem similar to theirs.
• What you would want to do or learn in the first months.

Write a cover letter of about 250 words for the job ad below. Use my notes for the reasons and the example; don't add any experience that is not in my notes or CV. Tone: confident and plain, no clichés like "I am writing to express my interest". Notes: [paste] Job ad: [paste]

The email writer is handy for the short message that goes with an application, or for a thank-you note after an interview.

Step 4: practise interviews

Llama makes a tireless mock interviewer. Set it up like this:

Play the hiring manager for the job ad below. Interview me with
6 questions, one per message: some about past situations, some about
the work itself. Wait for my reply each time, then tell me in two or
three lines what convinced you, what sounded vague and how you would
restructure it. At the end, name the two areas I should practise
most. Job ad: [paste]

Type your answers as you would say them. For behavioural questions ("Tell me about a time when…"), ask it to check your answer against the STAR structure: situation, task, action, result. Then say the improved version out loud. Reading feedback is not the same as answering under pressure.

You can also ask for the questions you hope nobody will ask: a gap in your CV, a job you left quickly, a change of career. Prepare a calm, honest answer for each one.

Step 5: research and networking messages

Llama can explain what a role or an industry typically involves, based on its training data, which helps you prepare good questions. It cannot tell you what a specific company is doing now, its recent news or what it is like to work there. Use the company's own website, recent articles and people who work there for that. Once you have collected that material, paste it in and ask Llama to turn it into five thoughtful questions to ask at the end of your interview.

For networking, ask for short and specific messages:

Write a LinkedIn message under 80 words to a former colleague who now works at [company]. I'd like 15 minutes to ask about the team's work, not a referral. Friendly, no flattery.

Salary and offers

Llama does not know current salaries in your city or field, and any figure it gives is a guess based on old data. Get numbers from up-to-date salary surveys, job ads that publish pay ranges and people who do the job. Where Llama helps is the conversation itself: "Help me write a polite reply to this offer, asking whether there is flexibility on the base salary because [your reasons]." You can then rehearse the call as a role-play, with Llama playing the hiring manager.

What to keep out of the chat

• ID numbers, your full date of birth, home address and phone number. None of the tasks above needs them.
• Your references' contact details and other people's personal data.
• Confidential information about your current employer.

More in AI privacy and safety. For ready-made prompts, browse the career prompts; for better prompting in general, read how to write prompts. Applying in another language? The translator is set up for that.

A simple weekly routine

1. Collect the ads you want to apply to.
2. Decode each one with the Step 1 prompt and drop the ones that do not fit.
3. Tailor your CV from the master version and draft the letter.
4. Edit both yourself, then run them through the grammar checker.
5. Do one mock interview session for the role you most want.

Before anything goes out, read it once more with fresh eyes and ask yourself three questions. Is every claim true? Could I talk about each line for two minutes? Does it still sound like me? If the answer to any of them is no, fix it by hand. A short, honest application beats a polished one you cannot defend in the interview.

FAQ

Will recruiters know I used AI?

They may notice generic phrasing. Edited, specific text built from your real experience reads like you, because it is about you. Some employers ask applicants about AI use; if they do, answer honestly.

Can Llama find job openings for me?

No. It cannot search the internet or job boards. Find openings yourself, then paste an ad into the chat to analyse it.

Can it translate my CV into another language?

Yes. Ask a native speaker to check the result before you send it, especially job titles and qualification names, which often have fixed local equivalents.

Is it all right to let AI write my whole application?

It is your application, so it is your decision, but you will be interviewed on everything in it. Use Llama for structure and wording, and make sure every claim is true and something you can explain in your own words.

Which Llama model should I use for applications?

The default, Llama 4 Maverick, is a good starting point for tailoring and letters, and it can read a screenshot of a job ad. For a second version of a letter, try Llama 3.3 70B.