End-to-end sentiment analysis system using DistilBERT with data cleaning, fine-tuning, accuracy evaluation, and real-time prediction support. Focused on efficiency, performance, and ease of deployment.

Built an end-to-end NLP service that flags potentially manipulative language in text for content moderation workflows. Fine-tuned a transformer model and deployed it behind a FastAPI inference API with a Streamlit web interface. The system demonstrates production-style ML deployment, model inference, and API-driven integration.

Fine-tuned DistilBERT for binary sentiment analysis on IMDB movie reviews with token-level interpretability using LayerIntegratedGradients

Successfully developed a multiclass text classification model by fine-tuning pretrained DistilBERT transformer model to classify various distinct types of luxury apparels into their respective categories i.e. pants, accessories, underwear, shoes, etc.

Code for a comparative analysis of the performance of fine-tuned transformer models on climate change data. The transformer models used were BERT, DistilBERT and RoBERTa.

The project involves developing a proof-of-concept system for classifying financial excerpts into predefined categories using Natural Language Processing (NLP) techniques.

Successfully established a multiclass text classification model by fine-tuning pretrained DistilBERT transformer model to classify several distinct types of mental health statuses such as anxiety, stress, personality disorder, etc. with an accuracy of 77%.

In this project we have tried to do multi-label hate-speech classification in Bengali and Hindi language using fill-mask transformer models.