career-ops-india

AI-powered job search pipeline for the Indian market

Works with Claude Code and Gemini CLI (free).
Scan 60+ Indian company career pages, evaluate fit with a structured rubric,
generate tailored PDFs, track your pipeline, prep for interviews.

Built for early-career folks targeting data, analytics, and product roles in India.

Inspired by career-ops by Santiago — rebuilt in full for the Indian market.

What it does

How it works

Most tools make you scrape job boards and get blocked. This doesn't.

Step 1 — Scanner hits ATS APIs directly

Companies like Razorpay, CRED, Zepto, Postman, and 56 others use Greenhouse, Lever, or Ashby as their applicant tracking system. All three expose clean JSON APIs. The scanner queries them directly — no browser, no CAPTCHA, no getting blocked.

node scripts/scan.mjs
→ Razorpay(2) CRED(1) Postman(3) BrowserStack(1) ...
→ 23 matching jobs found. Saved to data/scan_results.json.

Step 2 — AI evaluates each job against your profile

Open Claude Code or Gemini CLI in this folder. The AI has already read your CV, your target salary, and your skills. Paste a URL or run /batch on the scan results.

/evaluate https://boards.greenhouse.io/razorpay/jobs/6123456

[B] Razorpay — Data Analyst
Score: 4.1/5.0 | Bangalore | 14–18 LPA

✅ Strong: Python, SQL, pandas, ETL, Metabase — direct matches
⚠️  Partial: dbt mentioned, not in your CV
❌ Gap: 2+ years required, you have 1.5

Salary: Within target range. In-hand ~₹85–95K/month.
Recommendation: APPLY WITH TWEAKS

Step 3 — Tailored resume PDF per job

/pdf razorpay data analyst
→ Rewrites your bullets, injects 5 JD keywords, generates PDF
→ Saved: output/anoj-sk_razorpay_data-analyst_2025-06-01.pdf

Step 4 — Track, prep, negotiate

Everything stays local. Your pipeline, CV, and reports are gitignored — they never leave your machine.

Setup

Requirements

• Node.js 18+ → nodejs.org
• Claude Code or Gemini CLI (free) — at least one
• Git

Install

git clone https://github.com/AnojSKunte/career-ops-india
cd career-ops-india
npm install
npm run doctor        # checks everything is ready

Configure

cp config/profile.example.yml config/profile.yml

Edit config/profile.yml — add your name, skills, target roles, salary range, locations.

Then edit cv.md — replace the placeholder text with your real experience.
The more specific and quantified your CV, the better every evaluation will be.

First run

npm run scan

No AI key needed for scanning — just Node.js. You'll see matching jobs from 60+ companies in about 30 seconds.

Then open your AI CLI:

# Claude Code
claude

# Gemini CLI
gemini

And run:

/evaluate [any URL from the scan results]

Slash commands

All commands work identically in Claude Code and Gemini CLI.

/evaluate [URL or paste JD text]    Full A–F evaluation, 6-block report
/scan                               Re-scan companies, show new listings
/batch                              Evaluate multiple jobs at once
/pdf [company] [role]               Generate tailored resume PDF
/tracker                            Log an update to your pipeline
/pipeline                           View your full application dashboard
/prep [company] [role]              Interview preparation
/contact                            Write outreach messages
/negotiate                          Salary negotiation scripts

npm scripts

npm run scan          # Scan 60+ company ATS APIs for matching jobs
npm run doctor        # Check setup: Node version, files, CLI, Puppeteer
npm run dashboard     # Open visual pipeline dashboard in browser
npm run liveness      # Check if job application links are still live
npm run verify        # Validate data/pipeline.json integrity
npm run dedup         # Remove duplicate pipeline entries
npm run sync-check    # Check cv.md and profile.yml are in sync
npm run pdf           # Generate PDF (usually called via /pdf in AI CLI)

Company coverage

256 Indian companies pre-configured across 3 ATS systems:

To add a company: find their careers URL, identify the ATS from the domain, add a 4-line entry to portals/india.yml. That's it.

Sources covered

ATS companies (Greenhouse / Lever / Ashby) — npm run scan

The cleanest, most reliable source. Companies post jobs directly to their ATS. The scanner hits the JSON APIs — no browser, no CAPTCHA, instant results.

LinkedIn — npm run scan:linkedin

Uses LinkedIn's public guest API (/jobs-guest/). The same endpoint Google uses to index jobs. No login. No account. No cookie management. Parses HTML card responses. Adds 2–5s delay between requests to stay within rate limits.

npm run scan:linkedin
# Searches 10+ query variations across your target roles
# Adds results to data/scan_results.json
# Takes ~5 minutes

Naukri — npm run scan:naukri

Naukri has no public API. This uses Playwright to simulate a real browser. Slow by design — adds 5–10s delays to avoid detection.

# First time only:
npm install playwright
npx playwright install chromium

# Then:
npm run scan:naukri
# Takes 10–15 minutes
# Set HEADLESS=false in scripts/naukri.mjs if getting blocked

Note: Naukri's ToS prohibits scraping. Use for personal job search only, not for commercial data collection.

Why not just scrape Naukri / LinkedIn?

You could. There are two problems:

1. Reliability. Naukri uses Cloudflare + fingerprint detection. LinkedIn actively litigates against scrapers. Any scraper breaks within weeks when they update their frontend. You'd spend more time fixing the scraper than job hunting.

2. Signal quality. The companies worth targeting at your experience level — Razorpay, CRED, Zepto, Postman — all have ATS systems this scanner queries directly. Their listings are cleaner, more accurate, and have apply links that actually work.

Naukri is a volume play. ATS-first is a quality play. This tool optimizes for the latter.

Evaluation criteria (what A–F actually means)

Every job is scored across 10 dimensions:

Grade → Action:

• A (4.5–5.0): Apply today. Prioritize.
• B (4.0–4.4): Apply with tailored CV. Worth the effort.
• C (3.5–3.9): Apply if your pipeline is thin.
• D (3.0–3.4): Significant gaps. Skill-up first or reach out informally.
• F (<3.0): Don't apply. Time better spent elsewhere.

Privacy

Your personal data stays on your machine:

• cv.md — gitignored
• config/profile.yml — gitignored
• data/ — gitignored (pipeline, scan results)
• reports/ — gitignored (evaluation reports)
• output/ — gitignored (generated PDFs)

The repository only contains system files — modes, scripts, company list.
Nothing personal is ever committed or pushed.

File structure

career-ops-india/
├── CLAUDE.md                    ← AI brain for Claude Code (auto-loaded)
├── GEMINI.md                    ← AI brain for Gemini CLI (auto-loaded)
├── cv.md                        ← YOUR resume (gitignored — fill this in)
├── config/
│   ├── profile.example.yml      ← Template — copy to profile.yml
│   └── profile.yml              ← YOUR config (gitignored)
├── modes/                       ← AI instruction files (the core of the system)
│   ├── _shared.md               ← Scoring framework + Indian market context
│   ├── evaluate.md              ← Single job evaluation
│   ├── scan.md                  ← Portal scan + result interpretation
│   ├── pdf.md                   ← Tailored resume PDF generation
│   ├── batch.md                 ← Multi-job evaluation
│   ├── tracker.md               ← Pipeline tracking
│   ├── prep.md                  ← Interview preparation
│   ├── contact.md               ← Outreach messages
│   └── negotiate.md             ← Salary negotiation
├── portals/
│   └── india.yml                ← 60+ companies with ATS slugs (add more here)
├── scripts/
│   ├── scan.mjs                 ← ATS API scanner (no AI needed)
│   ├── generate-pdf.mjs         ← PDF generation via Puppeteer
│   ├── doctor.mjs               ← Setup health checker
│   ├── check-liveness.mjs       ← Dead link detector
│   ├── verify-pipeline.mjs      ← Pipeline data integrity
│   ├── dedup-tracker.mjs        ← Remove duplicate entries
│   ├── cv-sync-check.mjs        ← CV vs profile consistency check
│   └── open-dashboard.mjs       ← Opens browser dashboard
├── templates/
│   ├── cv-template.html         ← ATS-safe HTML template for PDF
│   └── dashboard.html           ← Pipeline visual dashboard
├── .claude/commands/            ← Slash commands for Claude Code
├── .gemini/commands/            ← Slash commands for Gemini CLI
├── .github/                     ← Issue templates, CI workflow
├── data/                        ← Your pipeline + scan results (gitignored)
├── reports/                     ← Evaluation reports (gitignored)
└── output/                      ← Generated PDFs (gitignored)

Roadmap

• ATS scanner — Greenhouse, Lever, Ashby (256 Indian companies)
• Job evaluation — A–F scoring, 10 dimensions, ghost job detection
• Batch evaluation mode
• Tailored PDF generation
• Pipeline tracker with conversion analytics
• Interview prep — SQL, Python, case study frameworks
• Outreach message templates
• Salary negotiation (India-specific)
• Browser pipeline dashboard
• Dead link detector
• Setup doctor + data integrity tools
• CI workflow
• Resume audit mode (cold CV review — no job needed)
• Skill gap analysis mode
• Referral finder + outreach mode
• Workday scraper (Flipkart, Swiggy, PhonePe)
• iimjobs / Instahyre integration
• Resume score without a specific job (general ATS optimization)
• Weekly email digest of new matches

Contributing

PRs welcome. Most useful contributions:

• New companies added to portals/india.yml — just 4 lines per company
• ATS slug corrections — slugs change when companies rebrand
• New modes — e.g., referral tracker, follow-up scheduler
• Bug fixes in scripts

See CONTRIBUTING.md for details.

License

MIT — use it, fork it, build on it.