keinplankarriere
Final presentation · AI Agents

Job hunting,
with an agent on your side

An AI career agent that finds the jobs, ranks them against your real experience, writes your CV and cover letter — and gets better every time you apply.
keinplankarriere.qantra.dev
keinplankarriere
The problem

Applying for jobs is a second job

Fragmented

The same roles are scattered — and duplicated — across LinkedIn, StepStone, Xing and the Arbeitsagentur.

Repetitive

Every serious application means rewriting the CV and cover letter. 30–60 minutes, per job.

Blind

No feedback loop: which skills are in demand? Which of your applications actually worked?

Result: people either send generic applications everywhere — or burn out tailoring a few.

keinplankarriere
The solution

One agent, the whole loop

Find
4 boards, full descriptions, duplicates merged
Match
ranked 0–100 against your real experience
Apply
tailored CV + cover letter, one click
Improve
learns from interviews & offers

Grounded in you

Upload your CV once — the AI turns it into a reviewed experience base. Every score and every document is built from what you've actually done. No fabrication, ever.

You stay in control

The agent prepares; you submit. No auto-apply, no stored job-board credentials, no ToS violations — a deliberate design choice.

keinplankarriere
Find

Four boards, one clean list

LinkedInStepStone XingArbeitsagentur
  • Search any query + location straight from the dashboard
  • Full job descriptions fetched per posting — the fuel for everything downstream
  • Skills auto-extracted (curated taxonomy, DE + EN)
  • Fuzzy dedup merges the same role across boards

Trends, for free

Because everything is structured, the Trends tab shows what the market wants right now: top skills, salary ranges, remote split, hottest locations and companies — with a one-paragraph AI summary.

keinplankarriere
Match

Ranked for you specifically

Hybrid scoring

Layer 1 — a deterministic 0–100 score (skills 45 · role 20 · location · remote · seniority · salary): instant and free for every job.

Layer 2 — the top candidates get an LLM re-score that reads your actual experiences and explains itself in plain language.

It learns from outcomes

Mark a job as interview or offer and its skills become a success signal: similar jobs rise in the ranking, and the AI scorer is told what worked. Plus smart caching — unchanged jobs never pay for a second LLM call.

Every score comes with a visual breakdown — you always see why a job ranks where it does.

keinplankarriere
Apply

From “interesting” to application-ready in one click

Tailored CV
your LaTeX résumé, every bullet rewritten for the job — nothing dropped
+
Cover letter
grounded in your experience, rendered to PDF
+
Apply kit
posting link + checklist + “mark as applied”

Make it yours

Your CV photo, your prompt add-ons (“emphasise leadership”, “max 250 words”), your own LaTeX template — all per user, all in Settings.

Always a valid PDF

Documents are compile-validated (Tectonic). If a tailored version fails, it falls back to your known-good template — generation never leaves you empty-handed.

keinplankarriere
Shipped this sprint

Public. Multi-user. Try it right now.

Cloudflare Tunnel

Served from a small home desktop, exposed safely over HTTPS — no open ports, auto-restarting.

A workspace per visitor

Cookie sessions: your jobs, experiences, documents and settings are yours alone. Idle workspaces auto-clean after 48 h.

Bring your own AI

Visitors paste their own LLM key (Groq is free, OpenRouter, …) — the host's credentials are never shared or exposed.

Fair-use queue

All heavy work runs through one serial worker with live “#2 in line” feedback — many users, one small machine, no meltdown.

https://keinplankarriere.qantra.dev

Scraping and rule-based ranking work with no key at all; the owner workspace sits behind an admin key.

keinplankarriere
Under the hood

Small stack, built to survive

The stack

FastAPI + SQLite · React + Vite · pure requests/BeautifulSoup scrapers · Tectonic for LaTeX→PDF · any OpenAI-compatible LLM (Kisski, Groq, OpenRouter, …) · one Docker compose, two containers.

Resilience, learned the hard way

  • Self-healing LLM — when the provider rotates models away, it re-picks an available one mid-request
  • Deterministic core: dedup, skills and base scoring need no LLM at all
  • Fallbacks everywhere: compile errors, rate limits, empty workspaces

Design principle: the LLM is reserved for judgment — parsing CVs, refining matches, writing documents. Everything else is plain, testable Python.

keinplankarriere
4 sprints · shipped & deployed

From zero to a public AI career agent

4
job boards, full descriptions
0–100
explainable match score
2 PDFs
CV + letter per job, one click
∞ users
isolated sessions + queue
Sprint 1: scrape & track → Sprint 2: match & profile → Sprint 3: generate & apply → Sprint 4: public, multi-user, self-improving.
keinplankarriere.qantra.dev — scan a job, get your CV
1 / 9
→ / space to advance · F for fullscreen