a job applicant’s struggle: why we’re doing this

Alex is staring at the screen, exhausted. Another morning, another two hours spent sifting through job postings that don’t make sense. One company wants ten years of experience for an entry-level role. Another claims ”remote work” but then buries in the fine print that it’s actually hybrid. And then there’s LinkedIn, which has somehow decided that, based on Alex’s five years in marketing, they’d make an excellent supply chain analyst.

Applying is worse. Each job requires a perfect CV, the right keywords, a tailored cover letter—yet after all this effort, Alex feels like their applications disappear into a void. No feedback, no clarity, no idea what’s wrong. Are they underqualified? Overqualified? Was the formatting off? Should they have mentioned that side project?

We’ve all been Alex at some point. The job search process is broken. And we’re fixing it.

what we’re building and why it matters

The idea is simple: a job application assistant powered by AI that actually helps. Not just another automation tool, not a mindless job board scraper, but something that guides applicants through the process, offers insights, and helps them position themselves effectively.

Right now, job seekers are stuck between bad recommendations, outdated advice, and an overwhelming process. Job boards throw out random suggestions, resume builders create the same bland CVs for everyone, and AI is already being used to reject applications—but no one is using it to help people apply smarter.

So what if we built something that didn’t just automate job applications but actually helped people understand them?

Something that:

• Finds jobs that make sense based on an applicant’s skills and experience.
• Explains why a job might be a good fit—not with a meaningless percentage score, but a real breakdown of strengths and challenges.
• Generates a tailored CV and cover letter that actually matches the company’s culture and expectations.
• Learns from past applications to refine and improve the process over time.

This isn’t about blasting out hundreds of applications. It’s about helping people apply smarter.

why we’re not jumping into AI just yet

It would be easy to jump straight into AI—after all, that’s what everyone is doing. But that’s exactly why most AI-powered job search tools don’t work.

Throwing AI at a problem without understanding the problem first leads to bad results. If we built an AI-powered application generator without a clear structure, it would just spit out generic, robotic nonsense.

Instead, we’re doing it right:

• First, we define the pain points—what part of job searching actually needs fixing?
• Then, we build a structured system—something that works before AI is even added.
• Finally, we integrate AI to enhance it, not replace it.

If an AI tool can’t function without AI, it was never designed well to begin with.

how we’re designing this without scaring non-techies

Building any application like this requires a few basic components, but let’s keep this non-scary:

1. A place to store user data – Think of it like a notebook where we keep your job preferences, past applications, and AI-generated CVs. That’s what our database (PostgreSQL) does.
2. A way to process information and make decisions – This is the brain of the system. When you request job matches or AI-generated cover letters, the brain needs to understand what you need and generate a response. We’re using FastAPI, which is just a fancy way of saying a system that quickly responds when you interact with the tool.
3. A way to find jobs – We need a job search assistant that doesn’t just rely on LinkedIn’s bad recommendations. The problem? Job postings are spread across the internet. Some companies allow easy access to their listings, but others don’t. So, we have two choices:
   • Use an official job API (if available).
   • Build our own scraper (like a research assistant that scans job boards and collects relevant postings).
   • Since many companies don’t provide an official API, we’ll likely build our own smart scraping tool using Scrapy or Selenium, which are the most common solutions for this.
4. A way to write applications that don’t sound robotic – AI can generate job applications, but we need to make sure it actually sounds human. When we add AI, we’ll be using GPT-based models, which are already used for things like ChatGPT and automated content writing. But to make sure it works well for job seekers, we’ll fine-tune it based on real user feedback.
5. A simple, easy-to-use interface – None of this tech matters if people can’t use it easily. The entire system will be accessible through a clean, web-based dashboard, where users can:
   • Upload their CV.
   • Browse job matches.
   • Generate custom cover letters.
   • Track which jobs they’ve applied for.

💡 Bottom line: We’re not reinventing the wheel—this setup is industry standard, used in thousands of web applications. We’re just applying it to job searching in a way that makes sense.

our first major milestone – now what?

This is it. The MVP definition is complete, the UX flow is mapped out, and the system design is locked in. We’ve hit our first major milestone, and now the real work begins.

So, what happens next?

1. We create detailed UI wireframes to visualise exactly how people will use this.
2. We start early backend development, focusing on job scraping and user data handling.
3. We test the first interactions before adding complexity.

But we’re not doing this in a vacuum. We’re building openly, transparently, in public.

And we need your input.

your thoughts? let’s talk.

If you’re a UX specialist, project manager, or system architect, we’d love to hear your thoughts.

• Are we missing anything?
• Any potential pitfalls we should prepare for?
• Do you see any flaws in how we’re approaching this?

And for those of you who’ve been frustrated by job searching—does this sound like something you’d actually use?

We’re not just building a tool—we’re challenging the way job applications work. If you’ve ever fought against broken hiring systems, if you’ve ever wondered why LinkedIn thinks you’d make a great forklift driver, if you’ve ever sent out 100 applications and gotten radio silence—we want your thoughts.

Tell us where we’re right. Tell us where we’re wrong. Let’s make job searching less terrible.