Aditya Singh

– Analysed over 8 years of sales data to identify key trends and deliver actionable business insights. – Conducted Data exploration to uncover correlations and identify useful features within the data to create useful features for forecast models. – Improved forecast accuracy through rigorous data cleaning, normalisation, feature engineering, and the inclusion of regional holidays and external variables in forecast generation. – Developed a forecasting pipeline integrating XGBoost, Prophet, SARIMAX, and stacking models to automatically select the lowest-MAPE model, improving overall prediction accuracy by 35%. – Validated forecasting models using comprehensive back-testing, cross-validation, and walk-forward prediction techniques. – Designed and implemented an automated ETL and forecasting pipeline integrating client databases, managed workflow orchestration with Airflow, and scheduled regular model retraining and report generation. – Created a dynamic model selection system combining external forecasting API results and internal ML models, applying performance-based weighting to select optimal forecasts, backed by automated back-testing, real-time monitoring dashboards, and deviation alerts. - Worked on experimental packages FastML (https://pypi.org/project/plantbrain-fastml), which is an automated ML library designed to accelerate the process of training, evaluating, and tuning of models with features model comparison, preprocessing, and hyperparameter optimization for quick experimentation and prototyping on data.

Responsible for tutoring Data structures and Algorithms to CSE sophomores.

Shortlisted and Selected for a minor degree program in Business Consulting offered by the Department of Management Studies (DoMS) at RGIPT In addition to my Major in CSE.

An Institute of National Importance setup under an act of parliament under the Ministry of Petroleum and Natural Gas in par with IIT's , NIT's and IIM's.