AutoJobApply PRO v2.0

Local Multi-Platform Agentic Job Search & AI Fit Ranker
NVIDIA NIM: Checking… 0 jobs seen
⚡
0
Ranked Discoveries
🎯
—
Top Fit Score
🌐
2 Active
LinkedIn + Internshala
👤
14 Skills
Candidate Profile
Discovery Query & AI Scorer
Fetch live job postings from selected platforms, deduplicate against seen registry, and score fit with NVIDIA NIM
Ready to run discovery.
Discovery Pipeline Stream
Live Ranked Job Postings
Click any row to reveal detailed AI reasoning, candidate alignment, and missing skills
💼
No ranked job results yet

Trigger a discovery search above, or verify your resume is ingested in the Candidate Profile tab.

Resume & Profile Ingestion
Upload your resume (.pdf or .txt) to automatically parse your skills, seniority, and preferences into profile.json
📄
Click or drop your resume file here
Supported formats: PDF, TXT, MD, LaTeX
NVIDIA NIM Connection
Configure your NVIDIA NIM API key for ultra-fast Llama-3.2 fit scoring
Checking…
Saved securely in .env on your local computer. Never committed or sent to external servers.
Candidate Scoring Rubric (config.json)
Define hard deal-breaker filters, score boosts, and location parameters
🔑 How to Generate Your Free NVIDIA NIM API Key
NVIDIA provides 1,000 free credits for Llama 3.2 and Meta models — zero GPU setup required
1,000 Free Credits
1
Visit the NVIDIA Model Catalog

Open build.nvidia.com in your browser.

2
Sign In or Create a Free Account

Click Sign In at the top right. You can sign in using your existing Google, GitHub account, or register a free NVIDIA developer profile.

3
Select Meta Llama 3.2 Model

Search for or click on meta/llama-3.2-11b-vision-instruct (or meta/llama3-70b-instruct).

4
Generate Your API Key

On the model detail page, click the green "Get API Key" button. Generate and copy your secret key starting with nvapi-....

5
Save in Dashboard Settings

Switch to the Rubric & AI Model Settings tab above, paste your key into the NVIDIA API Key field, and click Validate & Save. It will be stored safely in your local .env file (git-ignored).

⚙️ How AutoJobApply Works: Architecture & Pipeline
Local agentic discovery, full description enrichment, and intelligent candidate fit ranking
🌐
1. Multi-Platform Scraping

Discovers vacancies across Internshala (freshers, junior developer vacancies across India) and LinkedIn (global job search endpoint) matching your target role and location.

📑
2. Detail Scraping & Enrichment

For every job card discovered, AutoJobApply fetches the live detail page to extract the full description, required technical stack, and responsibilities.

🛡️
3. Deduplication Registry

Maintains a persistent ledger (seen_jobs.json). You never waste AI tokens re-evaluating or re-reading positions you already examined.

👤
4. Candidate Profiling

Extracts your skills, seniority level, and previous roles from your resume in documents/cv/ into a structured JSON benchmark (profile.json).

🧠
5. NVIDIA NIM Fit Scoring

Llama 3.2 grades each role from 0 to 100, strictly enforcing hard deal-breakers (e.g. compensation, language), applying score boosts, and identifying exact skill gaps.

🎯
6. Human-in-the-Loop Apply

Automated bot application submissions are dangerous and lead to portal blacklists. AutoJobApply sorts jobs by match score and lets you review AI reasoning before manually applying in 1 click.

💡 Platform Coverage & FAQs
Understanding platform differences, bot protection, and tips for best results
Why Internshala alongside LinkedIn?
Internshala provides fast, server-side rendered listings with 0 captchas and direct access to hundreds of active junior and fresher full-stack developer roles in India and Remote.
What about Naukri?
Naukri protects its web search endpoints with Akamai/Cloudflare bot shielding (HTTP 406: recaptcha required). Direct HTTP requests get blocked unless authenticated with active browser session cookies. Internshala and LinkedIn provide unblocked public access.
Is my personal data or API key safe?
Yes! Everything runs 100% locally on your machine. Your API key is stored in .env and your resume is stored in documents/cv/—both are strictly git-ignored and never shared.
How do I customize the ranking rules?
Go to the Rubric & AI Model Settings tab to edit hard deal-breakers (lines that trigger a 0 score) and score bonuses (e.g. +5 for remote-first or AI companies).