- Employee Salary Prediction System
- Internship Program: IBM SkillsBuild Virtual Internship
- Supported by: Edunet Foundation
- Platform: IBM SkillsBuild
- Duration: 6 Weeks (June 18, 2025 – July 30, 2025)
- Domain: Artificial Intelligence & Machine Learning
- My Name: Durga Prasad Papugani
The main goal of this project was to build a smart web application that could predict employee salaries based on different factors like experience, education, skills, and job role. I wanted to make it useful for HR professionals and companies to quickly analyze employee compensation using Machine Learning models and smart dashboards.
During the internship, I developed a complete Streamlit-based app with the following features:
- 🔍 Real-time employee salary prediction using multiple ML models
- 📄 Resume upload feature that auto-fills prediction fields
- 🧼 Complete preprocessing pipeline (cleaning, encoding, etc.)
- 📊 Interactive data dashboards for exploring insights
- 🔍 SHAP explainability for feature importance
- 📉 Residual plots, error analysis, and prediction distributions
- 📈 Model evaluation using R², RMSE, MAE, and CV
- 🗓 Trend charts and time-based analysis if date data is present
- 📊 Advanced statistical charts (Box, Violin, Outliers)
- 🧪 T-tests, correlation heatmaps, and hypothesis testing
- 📅 PDF report generation with predicted salary
- Linear Regression
- Random Forest Regressor
- XGBoost Regressor
- GridSearchCV for model tuning
- K-Fold Cross Validation for testing reliability
- Name: EMPLOYEE_DATASET
- Source: Kaggle
- Features Used: Department, Education, KPI metrics, Number of Workers, Target productivity (used as proxy for salary)
- Rows: 10000+..
- Preprocessing: Outlier removal, encoding, feature scaling, date parsing
- Frontend: Streamlit, CSS, HTML
- Backend/ML: Python, Pandas, Scikit-learn, XGBoost, SHAP
- Visualization: Matplotlib, Seaborn, Plotly
- PDF & Resume: PyMuPDF, FPDF, Regex
- End-to-end ML app from scratch (data to deployment)
- Built custom resume parser to extract user details
- Created PDF report generator with predictions
- Designed clean UI with animations and theme
- Model tuning and performance analytics
- Deployed the project to Streamlit Cloud
- Planned and designed the entire app structure
- Handled data cleaning and feature engineering
- Built all the ML models and evaluation logic
- Implemented resume parsing & PDF generation
- Created interactive dashboards with custom plots
- Managed styling, layout, and deployment
- Gained full experience in building ML web apps
- Improved understanding of real-world data workflows
- Learned Streamlit and PDF generation tools
- Practiced data visualization techniques
- Explored resume parsing using NLP/Regex
- Understood importance of UI/UX in ML apps
- Certificate: IBM SkillsBuild via Edunet Foundation
- Verified Badge: Issued through Credly
git clone https://github.com/DurgaPrasadPapugani/employee-salary-app.git- Use LLM (ChatGPT/Gemini) to analyze resume content
- Forecast future salary trends using time series models
- Support other domains beyond Indian job market
- Add login system, database, and session handling
Big thanks to Edunet Foundation and IBM SkillsBuild for giving me the opportunity to work on this amazing real-world AI project. It helped me grow technically and professionally.
