yureira / hr-machine-learning

HR Machine Learning

About . Prerequisites for Development . Directory Structure . Getting Started . Contributors .

---

About

This repository host Python based scripts for HR Machine Learning. It provides a skeleton for users to train models and hosting the model on fastapi to allow API calls to be made for predictions.

---

Directory Structure

│   .gitignore
│   README.md
│   requirements.in
│   requirements.txt
│   run.py
│
├───experiments
│       demo_notebook.ipynb
│
├───images
│       fastapi_demo.png
│       fastapi_output.png
│
├───src
│   │   config.py
│   │   main.py
│   │   utils.py
│   │
│   └───model
│       │   grid.pkl
│       └───models.py
│
└───tests
    └───test_utils.py

---

Prerequisites for Development

• conda environment
• Python version >= 3.9, <3.10
• Bash shell/unix terminal/git bash (for Windows)

---

Getting Started

Conda Environment:

# cd to preferred repository
git clone https://github.com/Bcpeh/hr-machine-learning.git

# creating a new conda environment
conda create -n myenv python=3.9

# installing required packages
pip install -r requirements.txt

# Running unit tests
pytest tests/test_utils.py

# Running fastapi
python run.py

Running Inference

Once your localhost is running visit http://localhost:2222/models

Example of sending an inference request via Swagger UI: Example of output: Example of sending an inference request via command line:

curl -X 'POST' \
  'http://localhost:2222/v1/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "enrollee_id": 12,
    "city_development_index": "0.827",
    "gender": "Male",
    "relevent_experience": "Has relevent experience",
    "enrolled_university": "Full time course",
    "education_level": "Graduate",
    "major_discipline": "STEM",
    "experience": "9",
    "company_size": "<10",
    "company_type": null,
    "last_new_job": "1",
    "training_hours": "21"
}

---

Contributors

If you have any questions please contact one of our contributors below

• Boon Chong Email