Pinned Loading

1. YouTube-AI-Analyzer YouTube-AI-Analyzer Public

   Analyze and compare YouTube video content with AI-generated video and comment summaries, sentiment analysis, real-time performance metrics, and natural language Q&A.

   JavaScript

2. Finetune-RoBERTa-Sentiment Finetune-RoBERTa-Sentiment Public

   Improves sentiment accuracy on YouTube comments by fine-tuning a multilingual XLM-RoBERTa model on 1M labeled comments. Achieves 80.1% accuracy (vs. 69.3% baseline).

   Jupyter Notebook

3. Portfolio-AI-Assistant Portfolio-AI-Assistant Public

   An Agentic RAG-based AI Assistant I built for my portfolio, answering detailed questions about projects, freelancing, skills, and timelines with accurate, context-aware responses.

   Jupyter Notebook 1

4. YouTube-Comments-Sentiment-Curation YouTube-Comments-Sentiment-Curation Public

   Curated a large-scale, multilingual dataset of over 1 million YouTube comments labeled with Positive, Neutral, and Negative sentiments across diverse topics and countries.

   Jupyter Notebook

5. Audiobook-Purchase-Predictor Audiobook-Purchase-Predictor Public

   A neural network model for predicting repeat audiobook purchases, using customer engagement data to support retention strategies. Achieved 92% accuracy with balanced data and optimized training.

   Jupyter Notebook

6. fraud-detection-system fraud-detection-system Public

   A machine learning system for detecting fraudulent financial transactions using 6M+ records. Tackles class imbalance with SMOTE, uses feature engineering and ensemble models (Random Forest) to impr…

   Jupyter Notebook