SPEC-EN/REV.07
Тема:
Mikhail Matveev
all projects
hh-job-tracker
hh-job-tracker · 2024 · public

hh-job-tracker.

Self-hosted HH.ru job-search tracker. Applicant funnel, ML match-score, invitation probability prediction. FastAPI + HTMX + SQLite.

§ stack · bill of materials
PostgreSQL Redis Django HTMX Docker Python Celery

What it is

A local self-hosted app for structured job hunting on hh.ru.
Scrapes vacancies and application data from hh.ru pages using a logged-in browser
session cookie, stores everything in SQLite, and serves it via FastAPI + Jinja2 + HTMX.

HH.ru's standard UI targets the mass applicant market — not someone looking for a
specific stack, salary, and work format. HH Job Tracker fixes that.

Key features

  • Applicant funnel — statuses: RESPONSE / INVITATION / INTERVIEW / DISCARD
  • Match-score 0–100 — vacancy ranking against your resume
  • Predict % — invitation probability via sklearn LogisticRegression trained on your
    actual history
  • Employer politeness index — % read responses, avg response time (HH Pro data)
  • Background sync — every 4–6 h, deduplication, auto-archiving
  • Resume profile import — import your stack and keywords directly from hh.ru
  • Filters and sorting — by stack, salary, format, status, date

Stack

Backend: Python, FastAPI, SQLAlchemy, SQLite
Frontend: Jinja2, HTMX, Vanilla JS
ML: scikit-learn (LogisticRegression)
Deploy: Docker, Makefile

§ end of sheet

Run it yourself?
Drop a message.