SANKET HINGANE

WOODSTREAM

• Collaborated with senior leadership to develop and implement advanced models used for Time series forecasting which forecast the demand of the products using libraries such as Pandas, NumPy, scikit-learn, and TensorFlow.
• Developed ARIMA, SARIMA, CNN, RNN, LSTM, Gluon TS, Fb prophet, XG Boost and machine learning algorithms for accurate forecasting which led to 30% proactive inventory management.
• Applied feature engineering techniques, model evaluation metrics, and visualization tools to enhance forecasting performance.

METRO - SYNTHETIC DATA GENERATOR

• Developed a Synthetic Data Generator to replicate real time data for the METRO team for studying different business use cases enabling robust training of predictive analysis models.
• Optimized synthetic data generation processes using parallel computing techniques, significantly reducing computation time and resource utilization for large-scale data generation tasks.
• Mitigated privacy concerns by generating synthetic data for cohort analytics and E-commerce data ensuring compliance with data protection regulations without compromising model accuracy.

• Build application using Python and streamlit which helps visualize data and take readings out of it
• Designed RESTful APIs that adhere to industry best practices, ensuring clear and intuitive endpoints, resource representations and versioning strategies
• Implemented secure authentication and authorization mechanism, including Oauth2, JWT and API keys to protect API endpoints and user data
• Conducted load testing and profiling to identify bottlenecks and optimize API performance, resulting in reduced response times and improved scalability
• Integrated API with third party services (Heroku and Digital Ocean), payment gateways ensuring seamless data exchange and functionality.