Debashruti Dian

I’m an aspiring Machine Learning Engineer with hands-on experience in predictive modelling, data analysis, and ML API deployment. I enjoy building end-to-end ML solutions—from data preprocessing and feature engineering to model optimization and evaluation.

As a Data Science Intern at Oasis Infobyte, I processed 10,000+ records to improve data quality and consistency. I developed regression and classification models reaching up to 88% accuracy, and improved performance by ~12% through feature engineering and hyperparameter tuning.

I also completed 100+ hours of AI/ML training at Samsung Innovation Campus, implementing supervised and unsupervised learning on structured datasets. I built neural network models using TensorFlow/Kera’s for classification and applied NLP techniques like tokenization and TF-IDF vectorization.

In my projects, I designed and deployed an end-to-end sales forecasting application and integrated the model with Fast API for real-time REST inference, boosting accuracy by ~15% using feature engineering and cross-validation. I also trained and optimized multiple classifiers for an Iris classification task, achieving 95%+ accuracy with confusion-matrix and classification-report evaluation.