AutoML Data Cleaning & Model Training System (ML Yantra)
Developed an AutoML system that automates data cleaning, preprocessing, feature engineering, and model training for structured datasets. The system transforms raw, unstructured data into model-ready formats, simulating real-world AI training pipelines. Worked extensively on handling missing values, outlier detection, encoding categorical variables, and feature selection to improve model performance. Designed workflows for automated model selection and evaluation, generating user-friendly reports. This project involved deep interaction with data quality, preprocessing strategies, and model optimization, which are critical components of AI training and annotation workflows. Also incorporated options for manual intervention, allowing users to understand and control each stage of the data preparation and modeling process.