JobFinder

A personal job portal that fetches remote, hybrid, and in-office jobs from 11+ boards, filters them to the roles / experience level / location you choose, refreshes daily, and shows everything in a dashboard with apply links, detected skills, and an optional OpenAI-powered resume tailoring tool.

Out of the box it's tuned for entry-level Data Engineer & AI/LLM/GenAI roles hiring in India, but everything is customizable in one config file — see Customizing your search.

Quick start

You only need Docker installed.

git clone https://github.com/k-Rohit/JobFinder.git
cd JobFinder
docker compose up -d --build

Open http://localhost:8787 — the first batch of jobs appears within a minute. That's it. Optionally:

• Tailor your resume to a job → open ⚙️ Settings, upload your resume and paste an OpenAI API key, then click "✨ Tailor resume" on any job.
• Pull Indeed + Naukri results → add a free JSearch (RapidAPI) key in ⚙️ Settings.
• Search different roles / country / cities → see Customizing your search.

Start the portal

Option A — Docker (recommended)

docker compose up -d --build

Then open http://localhost:8787. The jobs database and your uploaded resume persist in the jobfinder-data Docker volume across restarts.

Stop / restart / view logs:

docker compose down       # stop
docker compose up -d       # start again
docker compose logs -f     # follow logs

To preload API keys instead of using the dashboard ⚙️ Settings, create a .env file next to docker-compose.yml:

OPENAI_API_KEY=sk-...
JSEARCH_API_KEY=...
OPENAI_MODEL=gpt-4o-mini

Option B — run locally without Docker

./run.sh

To have it start automatically at login (so the daily refresh always runs):

cp com.jobfinder.portal.plist ~/Library/LaunchAgents/
launchctl load ~/Library/LaunchAgents/com.jobfinder.portal.plist

Customizing your search

Easiest — in the dashboard: open 🎯 Search profile at the top of the page. Add/rename roles (e.g. type "Data Analyst" with keywords data analyst, bi analyst), set your country, office cities, max experience, freshness, and favourite companies, then click 💾 Save profile. It applies immediately (hit ⟳ Refresh now for fresh results) — no files, no restart.

Advanced — config file: the same settings live in config.json, with a few extra knobs (region term lists, per-role search terms, ATS tokens). Copy the example and edit it:

cp config.example.json config.json   # then edit config.json

With Docker, put config.json inside the data volume so the container reads it:

docker compose cp config.json jobfinder:/data/config.json
docker compose restart

(Or set JOBFINDER_CONFIG=/path/to/config.json.) Key options:

Key	What it controls	Default
roles	Job roles to match: each has a label, title_keywords (decide if a title matches) and search_terms (sent to board search APIs)	Data Engineer, AI/LLM/GenAI
comfortable_years / max_experience_years	Experience window. Jobs asking more than the max are dropped; between comfortable and max are flagged "stretch"	1 / 3
onsite_cities	Cities where hybrid/in-office jobs are kept (remote is global)	Bangalore, Hyderabad, Pune
country + require_local_eligibility	Remote jobs must be open to this country; postings restricted elsewhere are dropped	India, true
max_age_days	Only show postings newer than this	7
favorite_companies	Companies tracked in the ⭐ Fav. companies tab — fetched from their ATS board (Lever/Greenhouse) where available plus a company-targeted LinkedIn search	Meesho, Swiggy, Zomato, Blinkit, Zepto, Myntra, Urban Company

The ⭐ Fav. companies tab tracks specific employers' DE/AI openings (using a broader role vocabulary — Data Scientist, Applied Scientist, SDE-Data, etc. — and showing senior roles too, since it's a company tracker). These bypass the freshness and onsite-city limits. Each company also gets a one-click careers link. Add a company with {"name": "...", "match": ["..."], "lever": "token"} (or "greenhouse": "token").

Changing roles, country, and onsite_cities automatically retargets the filters and the LinkedIn / JSearch / The Muse search queries. Set require_local_eligibility to false to keep all remote jobs regardless of region.

What it does

• Fetches from 11 sources (no keys needed): RemoteOK, We Work Remotely, Remotive, Jobicy, Arbeitnow, Himalayas, The Muse, Working Nomads, Jobspresso, NoDesk, and LinkedIn (via its public guest job-search endpoint, politely throttled).
• Indeed / Glassdoor: blocked behind Cloudflare, so direct fetching isn't possible — add a free JSearch (RapidAPI) key in ⚙️ Settings to pull their listings via Google-for-Jobs, or use the one-click search links.
• Location policy: every job must be open to candidates in India. Remote roles restricted to other regions (USA-only, Europe-only, …) are dropped; hybrid/in-office roles are kept only for Bangalore, Hyderabad and Pune. The Mode filter has a Remote · India only option, and remote jobs explicitly open to India get a 🇮🇳 badge.
• Filters for you: only Data Engineer / AI / LLM / GenAI / ML titles; drops Senior/Staff/Lead/Manager roles and anything asking 4+ years. Experience tags: Entry/Fresher, ≤ 1 yr, Unspecified, Stretch (2–3 yrs).
• Skill extraction: ~45 skills (Python, SQL, Spark, Airflow, dbt, AWS, LangChain, RAG, vector DBs, …) detected per posting and shown as tags.
• Fit score (0–100) ranks how suitable each job is for a fresher.
• Daily auto-refresh while the server runs, plus a "Refresh now" button.
• Track your pipeline: Save / Mark applied / Hide on every job, with stats.
• Resume tailoring: upload your resume (PDF/DOCX/TXT/MD) and add your OpenAI API key in ⚙️ Settings, then click "✨ Tailor resume" on any job. GPT rewrites your resume with that job's ATS keywords (never inventing experience) and you download the result as .docx or .md.

Configuration

Setting	How
OpenAI API key	⚙️ Settings on the dashboard, or export OPENAI_API_KEY=…
OpenAI model	export OPENAI_MODEL=gpt-4o (default gpt-4o-mini)
JSearch key (Indeed/Glassdoor)	⚙️ Settings, or export JSEARCH_API_KEY=…
Office-job cities	edit INDIA_HUBS in jobfinder/filters.py
Port	edit run.sh

Data lives in jobs.db (SQLite) and data/ (your resume). Only postings from the last 7 days are kept — older ones are rejected at fetch time and existing ones age out daily (saved/applied jobs are never pruned).