README.md

metadata

title: AI Recruitment Agent
emoji: ⚡
colorFrom: indigo
colorTo: green
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false

⚡ AI Recruitment Agent

A production-grade hybrid candidate matching pipeline using Groq LLM, Pinecone vector DB, and a Gradio UI.

Architecture

CSV Input → Stage 1: Normalize (Groq)
          → Stage 2: Embed + Match (Pinecone + SentenceTransformers) → Top 20
          → Stage 3: Deterministic Rerank (Groq) → Top 10
          → Stage 4: LLM Deep Review (Groq) → Top 5
          → Stage 5: Final Synthesis (Groq) → Shortlist

Setup (Local)

1. Install dependencies

pip install -r requirements.txt

2. Configure environment

cp .env.example .env
# Edit .env and fill in your API keys

3. Create Pinecone index

In your Pinecone console:

• Create an index named recruitment-index (or whatever you set in PINECONE_INDEX)
• Dimension: 384 for all-MiniLM-L6-v2, 1024 for BAAI/bge-m3
• Metric: cosine

4. Run

python app.py

Open http://localhost:7860

Setup (Hugging Face Spaces)

Do not commit a .env file. Instead, go to your Space → Settings → Repository Secrets and add:

Secret	Example value
GROQ_API_KEYS	gsk_xxx,gsk_yyy
GROQ_MODEL	llama3-70b-8192
PINECONE_API_KEY	pcsk_xxx
PINECONE_INDEX	recruitment-index
EMBEDDING_MODEL	all-MiniLM-L6-v2
STAGE2_TOP_K	20

CSV Format

Column	Variants accepted
name	full_name, candidate_name
email	email_address
skills	parsed_skills, technical_skills
experience	parsed_work_experience, years_of_experience
education	parsed_metadata_education
resume_text	parsed_summary, summary

Pipeline Stages

Stage	Method	Input	Output
1. Normalize	Groq LLM	All candidates	Structured features
2. Embed & Match	Pinecone + SentenceTransformers	All candidates	Top 20 by similarity
3. Rerank	Groq LLM (deterministic scoring)	Top 20	Top 10 with scores
4. Deep Review	Groq LLM	Top 5	Verdicts + signals
5. Final Synthesis	Groq LLM	Top 5 reviews	Final ranked shortlist