The work that prevents work (and why it matters)

Behind the scenes, weekly roundup building useresumate

Author

John Iseghohi
October 17, 2025

In partnership with

Guess what?

This week, I rolled out something super cool on useresumate.com that you might not spot in the UI or even think to add to your resume, but trust me, you'll totally feel the difference every time you use it!

I made sure the system never repeats the same task twice.

How awesome is that?

Useresumate empowers you to tailor your resume for each job application, ensuring you stand out by avoiding the common pitfall of using a generic resume for multiple positions.

Find out why 100K+ engineers read The Code twice a week

Staying behind on tech trends can be a career killer.

But let’s face it, no one has hours to spare every week trying to stay updated.

That’s why over 100,000 engineers at companies like Google, Meta, and Apple read The Code twice a week.

Here’s why it works:

  • No fluff, just signal – Learn the most important tech news delivered in just two short emails.

  • Supercharge your skills – Get access to top research papers and resources that give you an edge in the industry.

  • See the future first – Discover what’s next before it hits the mainstream, so you can lead, not follow.

The problem I solved

Here’s what used to happen, while I was testing the product:

  • Upload a job description, and the AI will analyze it, extracting key skills and identifying what truly matters.

  • Click "Analyze" again, and you'll receive the same input and output, consuming the same computational resources, while duplicate records accumulate in the database.

Not great, at all - this was a bug.

What changed

✅ Smart duplicate detection — The system recognises previously‑seen jobs and skips redundant analysis.

✅ Cleanup tools — New admin endpoints find and remove duplicate analysis that already slipped in.

✅ Better normalisation — Job text is normalised before comparison, so "Soft Skills" and "soft skills" don’t fork separate entries.

✅ Targeted database indexes — Purpose‑built indexes in our vector store make filtering and matching much faster.

Beyond deduping: precision in the matching engine

While refactoring the analysis pipeline, I upgraded how we match skills and responsibilities.

We (me & my AI models 😊) moved from string matching to understanding context. "Project management" and "managed cross‑functional projects" now resolve as related ideas, not unrelated strings.

I also centralised how we gather evidence — the concrete bullets and achievements that prove a skill.

Instead of scattered logic across endpoints, there’s a single, consistent path.

When we support your resume against a job, we pull the right evidence, every time.

The unsexy infrastructure that changes everything

Most of my week (late night: 6 pm - 12 midnight) was spent on the kind of work that rarely gets a launch video:

  • Reindex APIs to refresh data without full rebuilds

  • Safety checks around production signing keys to keep scale‑ups safe

  • Debug tooling to trace exactly what happens when your resume is processed

  • Schema refactors for flexibility when the real world doesn’t fit tidy categories

None of this is flashy.

None of it gets a feature announcement.

But this is the difference between a tool that works and a tool that scales.

Even though I’m not a software developer, I'm just a designer navigating my way through vibe-coding to build a real product that solves real pain-points.

When someone uploads 10 resumes and tests against 100 jobs, we don’t waste cycles on duplicates.

The matching engine runs fast and precisely.

If something breaks, we can trace and fix without guesswork.

What you’ll notice

  • Faster response times — Less redundant compute means quicker feedback

  • More accurate matches — Better context understanding in the engine

  • More reliable scaling — Built to handle growth without wobble

What’s next

The resume editor is receiving visible upgrades: more insightful suggestions, improved formatting, and additional guidance.

And it will sit on top of infrastructure that won’t let you down.

Want early access?

👉 Join the waitlist to try the more innovative matching engine and improved editor, plus get early‑support pricing.

Building this right means making the unglamorous investments that prevent wasted work — so the details that matter can shine.

Talk soon,

John

P.S. Have you used a tool that was almost perfect but did weird duplicate things or gave inconsistent results?
Hit reply and tell me. Those stories are my roadmap for what to fix next.