SHL Assessment Recommender

This is a SHL Assessment Recommender built using Streamlit and Google Gemini API. It provides an intelligent recommendation system for SHL assessments based on a job description or natural language query.

Key Features:

• Recommends SHL assessments based on job descriptions or queries.
• Uses Google Gemini API for enhancing job descriptions and extracting relevant metadata and test types.
• Returns recommended assessments in JSON format that includes relevant details like role, duration, test types, and more.
• Integrated with Sentence-BERT embeddings for semantic similarity-based recommendations.
• Provides a simple Streamlit web interface for users to input job descriptions and view recommendations.

---

Requirements:

Installation:

1. Clone the repository:

git clone https://github.com/lassmara/SHL
cd SHL-Assessment-Recommender

2. Install the required dependencies:

pip install -r requirements.txt

---

Configuration:

1. Google Gemini API Setup:

• Sign up for access to Google Gemini API (or use an alternative API for job description enhancement).
• Store the API key in the Streamlit secrets manager by adding the following entry in .streamlit/secrets.toml:

[GEMINI_API_KEY]
key = "your_gemini_api_key"

---

Running the App:

Locally:

1. Ensure that all required libraries are installed.
2. Run the Streamlit app:

streamlit run app.py

On Streamlit Cloud:

1. Push the code to your Streamlit Cloud app.
2. Add the Google Gemini API Key to the Streamlit Cloud secrets.
3. Launch your app on Streamlit Cloud and start querying for assessments.

---

How It Works:

1. User Input: Users input a job description or query into the text area.
2. Gemini API Integration: The input is processed by the Google Gemini API, which rewrites the job description into a structured format and extracts relevant metadata (such as job family, level, industry) and SHL test types (e.g., Cognitive, Personality).
3. Semantic Matching: Using Sentence-BERT embeddings, the job description is compared against SHL assessments stored in the dataset to calculate similarity scores.
4. Recommendations: The top K relevant assessments are returned as a JSON response. The recommendations include the following details:
   • Assessment Name and URL (linked to SHL catalog)
   • Remote Testing Support (Yes/No)
   • Adaptive/IRT Support (Yes/No)
   • Duration
   • Test Types
   • Description

---

Example JSON Output:

For the query "Hiring for a frontend engineer with JavaScript skills", the app might return the following recommendations:

[
  {
    "role": "[Java Developer](https://www.shl.com/assessment/java-developer)",
    "duration": "40.0 min",
    "test_types": "Cognitive, Technical",
    "remote_testing": "Yes",
    "adaptive_irt": "Yes",
    "description": "This assessment evaluates Java development skills."
  },
  {
    "role": "[Python Developer](https://www.shl.com/assessment/python-developer)",
    "duration": "60.0 min",
    "test_types": "Cognitive, Technical",
    "remote_testing": "No",
    "adaptive_irt": "No",
    "description": "This assessment evaluates Python development skills."
  }
]

---

Customization:

• You can modify the number of recommendations to display using the top_k parameter.
• The maximum duration for assessments can be filtered using the max_duration parameter.

---