📄 SHL Assessment Recommendation System – Solution Documentation

Author: [Veerababu Palepu]
Live UI Webapp: [https://shlassesmentrecommendationsystem.streamlit.app/
API Endpoint: https://shlassessementapiendpoint.onrender.com/recommend
Documentation Link: https://docs.google.com/document/d/1sxx3PIxQii_J7cgBx6FcNcF7dYggR5JBwr4CbwBIKOs/edit?usp=sharing

---

✅ Objective

Create a system to recommend the most relevant SHL assessments based on a job description or query. Deliver both:

• STEP 1: An interactive web application.
• STEP 2: A REST API endpoint that returns JSON recommendations.

---

🧠 Approach Summary

STEP 1 – Web UI (Streamlit)** Access Here: Live UI Webapp: [https://shlassesmentrecommendationsystem.streamlit.app/

• Built using Streamlit for a clean user interface.
• Users input a job description or role requirement.
• The system embeds both query and test descriptions using Google Gemini Embedding API (models/embedding-001).
• Computes cosine similarity to recommend the top 10 relevant SHL assessments.
• Displays results in a sortable, clickable table with URLs to the test pages.

STEP 2 – JSON API Endpoint (Render)**

• Created a FastAPI-based endpoint /recommend.
• Accepts a POST request with a query string.
• Returns a ranked list of the top 5 assessments in JSON.
• Deployed the backend using Render.com for public accessibility.

---

🛠️ Technologies Used

Component	Tool / Library
Embedding Model	Google Gemini embedding-001
Web App	Streamlit
REST API	FastAPI
Deployment	Streamlit Cloud, Render
Data Processing	pandas, NumPy, scikit-learn
Hosting	GitHub (for version control)

---

📦 API Usage

POST /recommend
URL: https://shlassessementapiendpoint.onrender.com/recommend

Request:

{
  "query": "Looking for cognitive ability and reasoning assessments"
}

Response:

{
  "recommendations": [
    {
      "name": "Logical Reasoning Test",
      "test_type": "Cognitive",
      "duration": "30 mins",
      "url": "https://example.com/logical-test"
    },
    ...
  ]
}

---

🔧 Here are 3 Easy Ways to Test API End Points:

Test the health endpoint in browser:👇

https://shlassessementapiendpoint.onrender.com/health

You should see: { "status": "healthy" }

1. Use Python Script (Recommended)

Run this locally in a .py file or Jupyter Notebook:

import requests

url = "https://shlassessementapiendpoint.onrender.com/recommend"
payload = {"query": "I am hiring for Java developers who can also collaborate effectively with my business teams. Looking for an assessment(s) that can be completed in 40 minutes."}

response = requests.post(url, json=payload)

print("Status:", response.status_code)
print("Results:\n", response.json())

---

2. Use Postman

• Method: POST
• URL: https://shlassessementapiendpoint.onrender.com/recommend
• Body → raw → JSON:

{
  "query": "Looking to hire mid-level professionals skilled in Python, SQL, and JavaScript."
}

---

3. Use curl in Terminal

curl -X POST https://shlassessementapiendpoint.onrender.com/recommend \
-H "Content-Type: application/json" \
-d '{"query": "Looking to hire analysts with strong cognitive and personality skills."}'

---