Articles

AI Recruiting

AI prompting do's + don't's: Writing a job description for software engineers

sj-niderost-headshot

SJ Niderost

Content Marketing Manager

Posted on

September 5, 2025

Most job descriptions could be posted by any company anywhere. They're generic and fail to capture what makes a role worth applying for. But AI can help you create job descriptions that make software engineers stop scrolling if you correctly prompt it.

The difference between a generic AI-generated job description and one that attracts top talent is context. AI agents are generalists who can write everything from poetry to strategy documents, but don't know what's crucial to your specific company or role. Without proper guidance, you'll get output that reads like every other posting on LinkedIn or Indeed.

The don'ts: How generic prompts fall flat

Here's what happens when you use a basic prompt like "Write a job description for a data scientist position":

"We are looking for a data scientist to join our team. You'll analyze data, build models, and provide insights to help make business decisions."

This description tells candidates what the job is, but not why they'd want it. Thousands of companies could post it, and a software engineer has no compelling reason to choose your opportunity over others.

The same problem applies to software engineering roles. A prompt like "Write a job description for a senior software engineer" produces requirements-focused content that fails to differentiate your company or connect technical work to meaningful outcomes.

The do's: Context-rich prompting for stronger results

Transform your AI output by providing comprehensive context about your company, product, and ideal candidate. Here's how to write an effective prompt:

  • Include company context: Share your company stage, size, mission, and what you're building. Instead of just "software engineer," try: "Write a job description for a senior software engineer at our healthcare AI startup (Series A, 40 employees) building an AI diagnostic tool for radiology."

  • Define the technical environment: Specify the tech stack, engineering challenges, and scale. Mention whether you're working with microservices, handling millions of users, or building greenfield products.

  • Describe impact: Connect the technical work to business outcomes or user impact. For a healthcare AI startup, highlight how the engineer's work directly affects patient outcomes and democratizes expert-level diagnostics.

  • Target the right audience: Identify who you're trying to attract. Are you targeting engineers leaving big tech who want more meaningful work? Those excited by early-stage challenges? Engineers passionate about healthcare technology?

“Context helps AI understand not what the job is, but why someone should want it. And so adding more context about your company's mission, product, target candidate, and motivations will actually attract the right talent.” - Einas Haddad, Head of AI Engineering at Gem

From generic to magnetic

The output from this detailed prompt leads with impact rather than requirements:

"Join [Company]’s mission to democratize expert-level diagnostic capabilities. As a Senior Software Engineer, you'll work with cutting-edge medical imaging datasets and collaborate directly with radiologists to build AI tools that directly impact patient outcomes."

This description stops people from scrolling because it connects technical challenges to meaningful work - exactly what many software engineers leaving big tech look for.

Advanced AI prompting techniques

Iterate and refine: Treat AI like a teammate. If your first result feels too formal, ask for a warmer or casual tone. If it's too long, request a more concise version focused on the most compelling aspects.

Show what to avoid: Be explicit about what you don't want. "Don't use buzzwords like 'rockstar' or 'ninja.' Focus on actual impact and growth opportunities instead of generic tech jargon."

Request multiple options: Ask for three versions — one more technical, one emphasizing company culture, and one highlighting career growth. This gives you options to test different approaches.

Use two-shot prompting: Share examples of job descriptions that have worked well for similar roles at your company. This teaches the AI your specific style and what resonates with your target candidates.

“Give the AI two good examples. Just give it examples. Like, what does good look like? So, beyond describing what you want, I want a great software engineer with this background. Give them two examples of the ideal output, and then ask the AI to generate more like them. Why does it work? [AI] learns from patterns. So you're teaching the model to think about what good looks like.” - Lacey Kim, Head of AI products at Gem

Why this matters

The engineers you want aren't just looking for technical challenges. They want to understand the impact of their work, the company's mission, and why this role represents the next step in their career. Give AI the complete picture of your company, product, and ideal candidate, and you'll create job descriptions that turn passive job seekers into active applicants.

Stop asking AI to write job descriptions and start asking it to craft compelling reasons for top engineers to join your mission.

Want more AI prompting tips? Watch our on-demand webinar featuring our very own AI experts.


Share

Request Demo Image

Get started today

See how Gem can help you hire with remarkable speed and efficiency