README.md

metadata

title: SHL Assessment Recommendation Engine
emoji: 🎯
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: mit
short_description: AI-powered SHL assessment recommendations using 2-Stage RAG

🎯 SHL Assessment Recommendation Engine

AI-powered recommendations for SHL assessments using a novel 2-Stage RAG Pipeline.

🏗️ Architecture

Query → SBERT Encoder → FAISS Index → Top-K → Cross-Encoder Reranker → Results

Component	Model	Purpose
Retriever	all-MiniLM-L6-v2	Fast semantic embedding
Index	FAISS IndexFlatIP	Efficient similarity search
Reranker	ms-marco-MiniLM-L-6-v2	Accurate pairwise scoring
Scoring	Hybrid	30% retrieval + 70% reranking

📊 Performance

Metric	Score	Interpretation
P@1	0.90	90% top-1 accuracy
MRR	0.95	First relevant at rank ~1
NDCG@5	0.944	Near-optimal ranking

🚀 Usage

1. Enter a job description or role requirements
2. Click "Get Recommendations"
3. View matching SHL assessments with scores

📚 Dataset

74 SHL assessments from the SHL Product Catalog including:

• Job Focused Assessments
• Cognitive Assessments
• Personality Assessments
• Skills & Simulations

---

Built for SHL Research Internship Assessment • December 2024